US20040095626A1 - Reference structures and reference structure enhanced tomography - Google Patents

Reference structures and reference structure enhanced tomography Download PDF

Info

Publication number
US20040095626A1
US20040095626A1 US10/388,480 US38848003A US2004095626A1 US 20040095626 A1 US20040095626 A1 US 20040095626A1 US 38848003 A US38848003 A US 38848003A US 2004095626 A1 US2004095626 A1 US 2004095626A1
Authority
US
United States
Prior art keywords
source
reference structure
radiation
resolver
dimensional reference
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Abandoned
Application number
US10/388,480
Inventor
David Brady
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
UNIVERSITY DUKE
Duke University
Original Assignee
Duke University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Duke University filed Critical Duke University
Priority to US10/388,480 priority Critical patent/US20040095626A1/en
Assigned to UNIVERSITY, DUKE reassignment UNIVERSITY, DUKE ASSIGNMENT OF ASSIGNORS INTEREST (SEE DOCUMENT FOR DETAILS). Assignors: BRADY, DAVID J.
Priority to PCT/US2003/011390 priority patent/WO2004032047A2/en
Priority to AU2003224956A priority patent/AU2003224956A1/en
Publication of US20040095626A1 publication Critical patent/US20040095626A1/en
Priority to US12/000,450 priority patent/US7912173B2/en
Abandoned legal-status Critical Current

Links

Images

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T11/002D [Two Dimensional] image generation
    • G06T11/003Reconstruction from projections, e.g. tomography
    • G06T11/005Specific pre-processing for tomographic reconstruction, e.g. calibration, source positioning, rebinning, scatter correction, retrospective gating
    • BPERFORMING OPERATIONS; TRANSPORTING
    • B33ADDITIVE MANUFACTURING TECHNOLOGY
    • B33YADDITIVE MANUFACTURING, i.e. MANUFACTURING OF THREE-DIMENSIONAL [3-D] OBJECTS BY ADDITIVE DEPOSITION, ADDITIVE AGGLOMERATION OR ADDITIVE LAYERING, e.g. BY 3-D PRINTING, STEREOLITHOGRAPHY OR SELECTIVE LASER SINTERING
    • B33Y80/00Products made by additive manufacturing

Definitions

  • the present invention relates generally to sensors and imaging systems, and more particularly to tomographic systems.
  • Tomography involves the measurement of a set of source projections from a source, and inverting the measurement to estimate source characteristics or variables.
  • the projections typically consist of electromagnetic or acoustic energy, but can also include any form of energy, which can be caused to propagate through a medium.
  • the most common method of tomographic imaging inverts a discrete set of projection measurements to estimate source distribution as a point wise projection in normal or hyperspectral space.
  • the inverted set of projections can yield an image of the interior of the medium through which the projections passed.
  • Tomography finds application in a wide range of detecting and imaging fields, such as medicine, seismology, and non-destructive testing. Tomography is a calculation intensive process done numerically by computers, and thus the process is also referred to as “computer tomography.”
  • An object of the invention is to solve at least the above problems and/or disadvantages and to provide at least the advantages described hereinafter.
  • the object of the present invention is to provide a reference structure tomography device and method that includes a multi-dimensional reference structure that is adapted to modulate electromagnetic fields along both longitudinal and transverse directions.
  • a source revolver comprising a multi-dimensional reference structure configured to modulate source radiation in a pre-determined manner to produce a reference structure modulated signal comprising a puerility of source space projections, where in the multi-dimensional reference structure comprises a longitudinal dimension, and an analyzer configured in accordance with said reference structure for resolving a source state from the reference structure modulated signal.
  • a source imager comprising a multi-dimensional reference structure configured to modulate source radiation to produce a reference structure modulated signal comprising a puerility of source space projections where in the multi-dimensional reference structure comprises a longitudinal dimension, a sensor array configured to detect the reference structure modulated signal, and a processor coupled to the sensor array and configured to derive source information from the detected reference structure modulated signal.
  • a tomographic apparatus comprising a source, a sensor, a multi-dimensional reference structure between the source and sensor for producing modulated radiation from source radiation, where in the modulated radiation is received by the sensor and comprises a puerility of a source space projections and where in the multi-dimensional reference structure comprises a longitudinal dimension, and an analyzer coupled to the sensor for analyzing the modulated radiation.
  • a method of enhancing a sensor comprising altering radiation received by the sensory by inserting a multi-dimensional reference structure into a path of the radiation, where in the altered radiation comprises a puerility of source space projections, and where in the multi-dimensional reference structure comprises a longitudinal dimension, and analyzing the sensors response to the altered radiation.
  • a sensory system comprising a multi-dimensional reference structure configured to receive unconditioned data radiated from a source and to produce condition data comprising a puerility of source space projections, where in the multi-dimensional reference structure comprises a longitudinal dimension, a sensory configured to receive the conditioned data, and an analyzer configured to invert the conditioned data received by the sensor.
  • a measurement system comprising a source configured to produce source modulated radiation, a multi-dimensional reference structure comprising a longitudinal dimension for receiving the source modulated radiation and producing reference modulated wave radiation comprising a puerility of source space projections, and an analyzer in communication with the reference modulated radiation for analyzing the reference modulated radiation.
  • a method of resolving a source state comprising modulating source radiation such that a puerility of source space projections are created in a parallel fashion, and resolving the source state using a puerility of source space projections.
  • FIG. 1 is a schematic representation of a reference structure tomography (RST) system, in accordance with one embodiment of the present invention
  • FIG. 2 is a schematic representation of a RST system, in accordance with another embodiment of the present invention.
  • FIG. 3 is a schematic representation of a RST system, in accordance with another embodiment of the present invention.
  • FIGS. 4 - 7 are schematic diagrams of various RST devices, in accordance with the present invention.
  • FIG. 8 is a schematic illustration of a basic structure for RST imaging, in accordance with the present invention.
  • FIG. 9 illustrates the ray geometry of FIG. 8.
  • FIG. 10 illustrates a point source geometry
  • FIGS. 11 - 15 are schematic diagrams illustrating a reference structure obscuring various point sources, in accordance with the present invention.
  • FIG. 16 is a schematic diagram of a RST system, in accordance with another embodiment of the present invention.
  • FIG. 17 illustrates a geometry for analysis of spatial scaling vs. resolution for two-dimensional RST, in accordance with the present invention
  • FIG. 18 illustrates a measurement circle
  • FIGS. 19 - 20 are three-dimensional schematic representations of RST systems, in accordance with the present invention.
  • FIG. 21 is a schematic illustration of a RST system that utilizes spherical geometry and obscuring sphere, in accordance with one embodiment of the present invention.
  • FIG. 22 is a schematic illustration of a reference structure formed around clear polygons in accordance with the present invention.
  • FIG. 23 b is a schematic representation of stacked reference structure, in accordance with the present invention.
  • FIGS. 24 - 29 are transmission graphs of the reference structure of FIG. 23;
  • FIG. 30 is a schematic diagram of a plane wave motion sensor, in accordance with the present invention.
  • FIG. 31 illustrates a connection matrix for the plane wave motion sensor of FIG. 30
  • FIGS. 32 - 33 are graphs that show the 3-dimensional output of the connection matrix of FIG. 31, in accordance with one embodiment of the present invention.
  • FIG. 34 is a schematic diagram of mask and sensor geometry, in accordance with one embodiment of the present invention.
  • FIG. 35 is a grid showing how measurement pixels may be numbered in the mask and sensor geometry of FIG. 34;
  • FIGS. 36 a - 36 c illustrate different views of a connectivity pattern for a reference structure, in accordance with the present invention
  • FIG. 37 is a graph of the singular value spectrum associated with the connectivity patterns of FIGS. 36 a - 36 c;
  • FIG. 38 illustrates a reference structure pipe with a length of 4.5 mm, in accordance with one embodiment of the present invention
  • FIG. 39 illustrates a reduced S matrix for the configuration of FIG. 38
  • FIG. 40 illustrates a connectivity pattern for the reference structure of FIG. 38
  • FIGS. 41 a - 41 k are cross-sectional schematic views of a set of masks, in order from an input side to an output side, which may be stacked or layered to form a reference structure in accordance with an embodiment of the present invention
  • FIGS. 42 - 43 are schematic diagrams that illustrate imaging of a 4 ⁇ 1 LED array using a reference structure, in accordance with an embodiment of the present invention
  • FIGS. 44 a - 44 o illustrate reference structure output patterns, in accordance with an embodiment of the present invention.
  • FIG. 45 illustrates a receiver pattern for a conventional motion sensor
  • FIG. 46 illustrates a receiver pattern for a motion sensor with a reference structure, in accordance with an embodiment of the present invention
  • FIG. 47 illustrates a receiver pattern for a motion detector with multiple reference structures, in accordance with an embodiment of the present invention
  • FIG. 48 is a perspective ghost view of a reference structure, in accordance with one embodiment of the present invention.
  • FIG. 49 is a photograph of a reference structure, in accordance with one embodiment of the present invention.
  • FIGS. 50 - 51 are photographs of output signals for the reference structure of FIG. 49.
  • the present invention utilizes the modulation of a selected physical property or set of properties of energy propagating between a source and a receiver in such a way that the receiver produces an enhanced measurement.
  • the utilization of this technique is hereinafter referred to as “Reference Structure Tomography” (RST).
  • a “source” can be either a generator, scatterer or absorber of propagating energy, and may be described by a source distribution function. Accordingly, a source can either emit energy, or can receive and transmit, obscure, reflect, refract, diffract or otherwise interact with propagating energy.
  • the propagating energy or projection is typically modeled as following a ray path through or from the source.
  • a source projection is a sum or integral of properties associated with a subset of the points in the source. The inversion of source projections can be done through numerical methods executed by a computer.
  • RST uses reference modulations in the space between sources and sensors to encode propagating energy, also referred to as fields, radiated or scattered from the sources.
  • the reference modulations are preferably achieved by using a reference structure.
  • RST preferably utilizes multi-dimensional reference structures to precondition projections for detection and/or subsequent inversion.
  • Various RST embodiments will be described below.
  • embedded detector RST is an embodiment in which the detector elements are contained within the reference structure.
  • Adaptive RST is an embodiment in which the reference structure changes or is changed in response to the state of the source or the measurement.
  • Impulse RST is an embodiment in which the projection is similar to an impulse response function, and the impulse response function of a RST system is the measurement produced by an impulse in the source space as a function of the position of the impulse.
  • a reference structure may consist of a multi-dimensional distribution structure both transverse and longitudinal to primary directions of propagation for the projection. If the radiation space is two-dimensional, such as waves propagating along a surface, the reference structure may also be two-dimensional. If the radiation space is three-dimensional such as waves propagating through a volume, the reference structure may also be three-dimensional.
  • a reference structure can be thought of as conditioning or encoding the radiation.
  • the goal of the encoding process is to condition source characteristics for computational estimation from sensor measurements.
  • the most typical objective is the estimation of the source distribution over a multi-dimensional space from measurements over a lower dimensional space. For example, one may seek to estimate the source density in a plane from measurements on a curve or one may seek to estimate the source density in a volume from measurements over a surface.
  • the simplest form of RST involves the estimation of sources from geometric projections modulated by obscurations.
  • a linear transformation space 14 receives source data 12 from a source space 10 .
  • the linear transformation space 14 transforms the source data 12 .
  • the transformed source data 16 propagates from the linear transformation space 14 and is received by a sensor space 18 .
  • This diagram is a generalized representation of a RST system, whereby the linear transformation space 14 may correspond to a reference structure, the source space 10 can correspond to a source, and the sensor space 18 can correspond to a sensor or sensor array.
  • the source data 12 and the transformed source data 16 correspond to propagating energy, such as electromagnetic radiation or acoustic waves, before and after interaction with the linear transformation space 14 , respectively.
  • FIGS. 2 and 3 are schematic representations of two examples of a RST system.
  • a source 20 is surrounded by a reference structure 21 , which consists of obstructions 22
  • the reference structure 21 is surrounded by sensors 24 .
  • the circular geometry of FIG. 2 shows how a source 20 can be completely surrounded by the reference structure 21 and the sensors 24 , suitable for a circular coordinate system.
  • FIG. 3 a radiating field 26 is received by a reference structure 28 , which is placed between the radiating field 26 and the detectors 30 .
  • the reference structure 28 consists of obstructions 22 .
  • the geometry of FIG. 3 has a transverse direction and longitudinal direction, and shows how a reference structure and detectors can be modeled in a rectilinear coordinate system.
  • a direct projection is a bundle of ray projections integrated across a range of angles.
  • a ray projection may be considered a visibility projection integrated as the angular range approaches zero.
  • a diffuse projection is a projection for points not distributed along a line, and a diffractive projection is a diffuse projection generated by diffractive wave propagation.
  • RST offers the advantage of extended depth of field and other data cube projections.
  • data cube originates in hyperspectral imaging, and means a two dimensional image viewed as a function wavelength or frequency forms a “data cube” consisting of a stack of images at different colors. Images in higher spaces, such as a three dimensional source viewed as a function of color, may form multi-dimensional or hyper data cubes. Generalizing to a higher level, one may view the data cube as the set of all spatially related measures of the state of a source.
  • the data cube refers to the cube describing the field intensity of a three dimensional source.
  • Extended depth of field consists of creating a 2D image from a 3D source by integrating line integrals through the source (a 3D source consists of a stack of images. An extended depth of field image smashes all of these images into a single plane.
  • Extended depth of field or slices of the data cube as a function of space or color are conventional projections of the data cube.
  • h would be a delta function along a line.
  • h would pick a color of the data cube.
  • An unconventional projection would be something other than the projection that picked a slice or a line of the data cube.
  • diffractive field propagation is propagation of the field according to a wave equation, i.e., normal propagation of an optical or acoustic field
  • spatial-spectral source is a source described by a data cube that includes both spatial and color axes.
  • a source is used herein to mean a set of vectors or functions that describe a source.
  • a source may be described by the amplitudes of impulse response functions sampled on a regular spatial grid.
  • a source may be described by a Fourier series, in which case the basis consists of sine and cosine functions.
  • a “non-standard basis” would be a description of a source using basis functions that are not regular sampling functions on a Euclidean space or that are not a simple Fourier basis.
  • RST systems may map the source space onto the measurement space so as to immediately produce unconventional projections of the source space.
  • Other examples include systems that project the spatio-spectral source data cube onto nonstandard basis.
  • RST enables reconstruction without scanning or with reduced scanning.
  • Related art computer tomography scans an illumination and/or detector system to serially acquire projections.
  • a reference structure sorts and creates projections such that complete data can be obtained in a single time step.
  • the number of measurements made for a given source size and resolution is approximately the same in both cases.
  • Another advantage of RST arises when the source can be represented by a set of basis functions with fewer components than the related art locally sampled basis. For example, if only one or a few source points are active, the source can be represented by a number of basis functions equal to the logarithm of the number of spatial resolution cells. In such cases, RST can reconstruct the source using many fewer measurements than conventional systems.
  • RST can be more efficient because it is a multiplex technique. Multiplexing enables one to sort and find sources in a way that is more efficient than measuring all possible source states. RST also offers the possibility of revealing more information about a measured space by increasing the complexity of the energy received by the sensors.
  • Multiplex sensing consists of measurements that depend jointly on multiple source variables.
  • a projection which depends on the state of a source at multiple points is a multiplex measurement.
  • Tomography relies on multiplex sensing and can be implemented using detector elements external to a reference structure that is integrated between the source and detector elements, or by using detector elements wholly or partially embedded in the reference structure.
  • RST may be implemented with static or time varying reference structures.
  • FIG. 4 is a schematic diagram of another embodiment of a RST system 100 .
  • the RST system of FIG. 4 includes a source 32 from which a source wave 34 emanates.
  • a reference structure 36 receives the source wave 34 , and a reference structure modulated wave 38 propagates from the reference structure 36 .
  • the reference structure modulated wave 38 is received by a sensor 40 .
  • the sensor 40 produces a signal in accordance with the reference structure modulated wave 38 and directs that signal to a processor 44 through a signal line 42 between the sensor 40 and the processor 44 .
  • the RST system 100 is configured so that the reference structure 36 receives a source wave 34 from the source 32 , and provides a reference structure modulated wave 38 which is received by the sensor 40 .
  • the reference structure 36 modifies the source wave 34 in accordance with an interaction between the source wave 34 and the reference structure 36 .
  • the reference structure modulated wave 38 is modulated based on a combination of the characteristics of the source 32 from which the source wave 34 originated and the reference structure 36 .
  • the processor 44 is configured to invert the response of the sensor 40 using a predetermined transformation function of the reference structure 36 to characterize the source 32 .
  • FIG. 5 is a schematic diagram of a RST system 200 with an active reference structure 54 , in accordance with another embodiment of the present invention.
  • the active reference structure 54 receives the source wave 52 and a reference structure modulated wave 56 propagates from the active reference structure 54 .
  • the active reference structure 54 also receives a signal from a reference structure driver 68 through a signal line 70 between the active reference structure 54 and the reference structure driver 68 .
  • the reference structure modulated wave 56 is received by an analyzer 58 .
  • the analyzer 58 includes a sensor 60 , which is connected to a processor 64 through a signal line 62 .
  • the analyzer 58 is coupled via a signal line 66 to the reference structure driver 68 .
  • the active reference structure 54 receives a feedback signal from the analyzer 58 and the reference structure driver 68 .
  • the active reference structure 54 can alter its characteristics in accordance with variations in the source wave 52 as detected by the sensor 60 from the reference structure modulated wave 56 .
  • the RST system 200 also includes a wave generator 46 , which generates an illumination wave 48 .
  • the illumination wave 48 is received by the source 50 and the source wave 52 then propagates from the source 50 .
  • a source wave 52 is not necessarily originally generated by the source 50 , but can be instead a wave which can either be generated by the source 50 , or can be received and reflected, transmitted, or otherwise propagated from the source 50 .
  • FIG. 6 is a schematic diagram of a RST system 300 , in accordance with another embodiment of the present invention, utilizing the wave generator 46 , illumination wave 48 , source 50 and source wave 52 of FIG. 6.
  • a point type reference structure 72 is used from which a reference structure modulated wave 56 propagates.
  • the reference structure modulated wave 56 is received by a plurality of detectors 60 , which form a sensor array 74 with sampling points corresponding to each of the detectors 60 .
  • the sensor array 74 has a signal line 76 coupled to a processor 64 .
  • the discrete sampling points (detectors 60 ) of the sensor array 74 of the RST system allow the reference structure modulated wave 56 to be discretely sampled.
  • FIG. 7 is a schematic diagram of a RST system 400 with a source wave modulator, in accordance with another embodiment of the present invention.
  • this embodiment includes a source wave modulator 88 from which a modulated illumination wave 82 emanates to be received by the source 50 .
  • a source wave 84 from the modulated illumination wave 82 propagates from the source 50 and is received by the reference structure 36 .
  • a reference structure modulated wave 86 then propagates from the reference structure 36 to be received by a sensor 40 .
  • the source wave modulator 88 is modulated by a modulator driver 80 through a signal line 78 . Accordingly, the source wave modulator 88 can modulate the illumination wave 48 , so that the source 50 receives a modulated illumination wave.
  • visible means detectable by the sensor and “visible projection” means that each sensor detects along the lines that are not obscured to it from the source.
  • computer tomography consists of the measurement and numerical inversion of ray projections, and can be enhanced by the inclusion of a reference structure.
  • a RST system may measure ray projections, but these projections will depend jointly on the source and the reference structure.
  • Multiplicative reference structures are structures which either completely obscure or completely transmit the energy of the ray projection.
  • FIG. 8 is a schematic illustration of the basic structure for a RST imaging system, for the case of ray projections.
  • a source 20 is surrounded by a reference structure 90
  • the reference structure 90 is surrounded by detectors 24 .
  • the detectors 24 receive energy from the source 20 , as represented by ray projections 92 , 94 , 96 .
  • T i ( ⁇ right arrow over ( ⁇ ) ⁇ , ⁇ right arrow over ( ⁇ ) ⁇ ) is the transmittance of the reference structure for rays incident at angle ⁇ right arrow over ( ⁇ ) ⁇ relative to a ray normal to the detector
  • ⁇ right arrow over ( ⁇ ) ⁇ is a transverse position on the detector surface
  • is a distance along the ray in direction i ⁇ right arrow over ( ⁇ ) ⁇
  • T i ( ⁇ right arrow over ( ⁇ ) ⁇ , ⁇ right arrow over ( ⁇ ) ⁇ ) is the x-ray transform of the reference structure (the Radon transform in 2D).
  • S( ⁇ ) is the state of the source described as a function of position in space (i.e. a source density function.)
  • S would be the source, e.g., if you look at a computer tomographic image of bone and tissue, S is the 3D function describing the intensity of the image as a function of position.
  • is a scalar parameter of integration.
  • An x-ray projection is the integral of S along a line through the source.
  • the set of all line integrals of S is the Radon transform. It is not quite the Radon transform in 3 and higher dimensions, but is related.
  • the Radon transform is invertible for S, which forms the basis for 2D computer tomography.
  • Computer tomography often reconstructs sources in 2D slices using Radon transforms. One obtains these transformations in computer tomography by scanning the illumination to get projections of the source along lines.
  • Reference structure tomography uses the reference structure to get discrete groups of line or other projections without scanning, and “integration along the ray extent” means integration from one end of the ray to the other.
  • the ray extent would be the length of the ray or the space spanned by it, and the state of the ith detector is the amplitude of the signal it is currently receiving.
  • FIG. 9 illustrates the ray geometry of FIG. 8.
  • a detector 106 has dimensions P 1 ⁇ P 2 and receives ray projections 98 , 100 , and 102 .
  • the normal vector to the surface of the detector 106 is 104 .
  • the angle between the normal vector 104 and ray projection 98 is ⁇ 2
  • the angle between the normal vector 104 and ray projection 102 is ⁇ 1 .
  • the reference structure may reduce the number of time steps and number of detectors needed to characterize the source for purposes of numerical inversion. Note that the measurements m i constitute a discrete linear transformation of the source state and that this transformation is invertible.
  • An embodiment of visible projection RST may include a source consisting of a 2 ⁇ 2 array of point sources of light.
  • the geometry of such point sources is illustrated in FIG. 10, and are labeled S 11 , S 12 , S 21 , S 22 , corresponding to their position in an array.
  • the sources may be measured, for example, by a detector array/reference structure combination, such as the one illustrated in FIGS. 11 - 15 .
  • a reference structure containing obstructions 111 lies between the point sources S 11 , S 12 , S 21 , S 22 and detectors 107 , 108 , 109 and 110 .
  • FIG. 12 the reference structure of FIG. 11 obscures all point sources from the first detector 107 except for point source S 12 because it lies outside the obscured area 112 .
  • FIG. 13 shows how the second detector 108 can detect only source points S 11 and S 21 because they lie outside the obscured area 114 .
  • FIG. 14 shows how the third detector 109 can detect only points S 12 and S 22 because they lie outside the obscured area 116 .
  • FIG. 15 shows how the fourth detector 110 can detect S 12 , S 21 and S 22 , but not S 11 , because they lie outside the obscured area 118 .
  • the reference structure is designed such that T is nonsingular, than this transformation may be inverted to estimate the source from the measurements.
  • the process of inversion consists of multiplying the measurement vector by the inversion matrix. This is called “algebraic inversion” and is always possible. In practice, more efficient inversion algorithms may be available and preferred.
  • the matrix T ⁇ 1 is the linear transformation
  • a “source descriptor” describes the actual state of the source, however, a measurement is what is actually measured.
  • a linear inversion or some other mathematical algorithm is used to estimate source descriptors or functions of source descriptors from the measurement. These estimates are what the source calls “source parameters,” also known as “source estimates” or “estimation parameters”.
  • the source descriptor could be “the state of the source is heads” or the “state of the source is tails”.
  • the measurement could be a picture of the coin. Based on the measurement, one could estimate whether or not the state of the source is heads or tails. If the image is blurry or noisy, this estimate may be subject to error.
  • the process of RST consists of using a reference structure to code the source radiation such that measurements can be used to compute estimations of source descriptors.
  • a linear transformation such as the matrix r described above is “well-conditioned” if it has an inverse matrix T ⁇ circumflex over ( ) ⁇ 1. This is the case if T is nonsingular and if the ratio of the largest eigenvalue or singular value of T to the smallest is not large.
  • the matrix in the above example is well conditioned.
  • Multiplicative diffuse RST takes advantage of wave phenomena such as diffraction, refraction, scattering, diffusion, phase, frequency, etc. Visible projection RST neglects diffraction. Accordingly, under visible projections, a measurement point sums all source points based on their visibility from the measurement point. The visibility is proportional to the density of the ray through the reference structure from the measurement point to the source point. Under diffuse RST, radiation from source points may diffract around obstructions. In this case, the state of a measurement may depend on source points that are not visible from a source point.
  • visible projection corresponds to looking at an object through a forest.
  • Light from the object may be sensed only if the object is unobscured by trees.
  • the object can be sensed even if it is obscured by the trees because sound waves diffract around the trees.
  • the signal diffracted by the trees still has some spatial sensitivity, i.e., the object may be detected through the trees differently depending on the object's and the detector's position relative to the trees).
  • This diffuse spatial pattern to find or image the person through the trees is diffuse RST.
  • Diffuse RST does not function with rays, visible RST does. Neither diffuse not visible RST are likely to rely on Radon transforms or any other well known transformation. They can be designed to rely on some well-known mappings, which we have done in the case of the motion sensor reported below, but in general the mapping and its inverse may be an arbitrary linear transformation, such as transformations T shown above. A reference structure relying on such wave characteristics will generate more complicated mappings between the source and the measurement, and can thus yield more information about the source.
  • a reference structure 122 is disposed between a source 124 and a detector 120 .
  • a receiver pattern 126 is shown superimposed over the source 124 . It should be noted that the receiver pattern 126 will not have rectilinear boundaries in the diffuse case.
  • FIG. 16 is a schematic diagram of a RST system in which receiver patterns for diffuse propagation, analogous to those shown above for the 2 ⁇ 2 source of FIGS. 10 - 15 , are shown.
  • ⁇ right arrow over (r) ⁇ is a position variable over the source space 124 .
  • ⁇ right arrow over ( ⁇ ) ⁇ is used as in Equation 1 to refer to position on the sensor.
  • the sensor is a finite area device that integrates the field over the area to produce a discrete number.
  • T i ( ⁇ right arrow over (r) ⁇ , ⁇ right arrow over ( ⁇ ) ⁇ ) is the receiver pattern 126 for position ⁇ right arrow over ( ⁇ ) ⁇ on the ith detector.
  • T i ( ⁇ right arrow over (r) ⁇ , ⁇ right arrow over ( ⁇ ) ⁇ ) is a probability that the source signal radiated from point ⁇ right arrow over (r) ⁇ is detected by the ith detector.
  • the order of integration in Eq. 2 is over the source variable ⁇ right arrow over (r) ⁇ on the inner integral followed by the detector position ⁇ right arrow over ( ⁇ ) ⁇ on the outer integral.
  • the order of integration can be flipped, however and one achieves a simpler and mote intuitive equation by performing the integral over ⁇ right arrow over ( ⁇ ) ⁇ first, which yields
  • receiver pattern 126 may interpret receiver pattern 126 to represent this average receiver pattern on the ith detector without loss of generality.
  • a mapping is a transformation from a distribution on one space to a distribution on another space.
  • Eq. (4) is a mapping from the source distribution defined over ⁇ right arrow over ( ⁇ ) ⁇ in the modified Eq. (4) to the sensor state defined over the discrete positions indexed by i.
  • Eq. (4) could describe either visible or diffuse RST, the difference between the two cases is that visible RST assumes that the receiver pattern is based on the “ray visibility” through the reference structure, diffuse RST allows the field radiated by the reference structure to diffuse through the structure.
  • the source field may consist of sound waves. The detectors can hear the sound even though the source may be visibly obscured because sound does not propagate in straight lines. However, the reference structure can still modulate the propagation of the sound waves such that a receiver pattern is produced and the source can be estimated.
  • T( ⁇ right arrow over ( ⁇ ) ⁇ , ⁇ right arrow over ( ⁇ ) ⁇ ) in Eq. 1 is not the same function as T i ( ⁇ right arrow over (r) ⁇ , ⁇ right arrow over ( ⁇ ) ⁇ ) in Eq. (4).
  • T( ⁇ right arrow over ( ⁇ ) ⁇ , ⁇ right arrow over ( ⁇ ) ⁇ ) is the visibility from the sensor point ⁇ right arrow over ( ⁇ ) ⁇ through the reference structure in direction ⁇ right arrow over ( ⁇ ) ⁇ .
  • calculation of the mapping from the source to the sensors requires detailed analysis of field propagation through the reference structure.
  • T i ( ⁇ right arrow over (r) ⁇ , ⁇ right arrow over ( ⁇ ) ⁇ ) describes the results of this analysis between the source point ⁇ right arrow over (r) ⁇ and the sensor point ⁇ right arrow over ( ⁇ ) ⁇ , but we do not consider the details of this diffractive and refractive analysis here.
  • Eq. (1) is a triple integral because we use a simple algorithm for relating the reference structure to the mapping between the source and the sensor in visible RST (the visibility through the reference structure determines the mapping).
  • Eq. (4) is only a double integral because we have not related the receiver pattern to the physical characteristics of the reference structure.
  • the receiver pattern 126 shown on the source 124 is a probability density where signals or rays emitted from the corresponding source points are sensed by the detector 120 after modulation by the reference structure 122 .
  • the receiver pattern 126 is nonconvex (e.g., along lines through the receiver pattern one encounters multiple disjoint regions of high and low probability of detection.
  • a convex region is a region such that all lines between points contained in the region are also contained in the region.
  • a connected region is a region such that there exists at least one path between any two points in the region which is also in the region.
  • a circular receiver pattern is both convex and connected. The cone associated with the receiver pattern for a pixel in a focal system is nonconvex but is connected.)
  • Nonconnected receiver patterns may be desirable because they enable high-frequency and high resolution sensing without localizing the receiver pattern to a point.
  • Non-monotonic modulation is a modulating signal that contains multiple local minima and maxima (it does not increase or decrease monotonically).
  • a lens does not modulate a wavefield in depth and modulates the optical path length monotonically up to some maximum and then monotonically back down to a minimum.
  • a reference structure will include multiple minima and maxima in the modulating refractive, diffractive or absorptive structure along lines through the reference structure.
  • Tomographic imaging systems can be based on particular models for the source/detector mapping. Under the multiplicative RST model discussed above, it is assumed that projections consist of a linear modulation on propagation through the source and the reference structure individually, but that the reference structure modulates the source signal multiplicatively. This model is appropriate when the reference structure strongly modulates the source signal, as when the reference structure consists of obscuring structures or apertures in an opaque mask. This is the case in the 2 ⁇ 2 example of FIGS. 11 - 15 .
  • t( ⁇ right arrow over ( ⁇ ) ⁇ , ⁇ right arrow over ( ⁇ ) ⁇ ) is the signal transmitted by the combination of the source and reference structure to the measurement point ⁇ right arrow over ( ⁇ ) ⁇ along the ray direction ⁇ right arrow over ( ⁇ ) ⁇ .
  • T ⁇ ⁇ ( ⁇ ⁇ , ⁇ ⁇ ) exp ⁇ ( - ⁇ ray extent ⁇ [ T ⁇ ( ⁇ ⁇ + ⁇ ⁇ ⁇ i ⁇ ⁇ ) ] ⁇ ⁇ ⁇ ) ( 9 )
  • volume hologram based on 3D absorption or phase modulation.
  • 3D imaging using volume holograms is condensed in, for example: G. Barbastathis, M. Balberg, and D. J. Brady, “Confocal microscopy with a volume holographic filter,” Optics Letters , vol. 24, pp. 811-813, 1999; G. Barbastathis and D. J. Brady, “Multi-dimensional tomographic imaging using volume holography,” Proceeding of the Ieee , vol. 87, pp. 2098-2120, 1999; or G. Barbastathis and D. J.
  • an additive reference structure where the reference structure weakly modulates the source signal, may be preferred. Weak modulation is attractive when the additive mapping, represented by Eq. (12) below is desired. This mapping is attractive when the detector signal is proportional to the intensity of the radiation field, as represented in Eq. (13) below.
  • the reference structure signal will add to the source, yielding the mapping: m i ⁇ ⁇ detector area ⁇ ⁇ detector acceptance angle ⁇ ⁇ ⁇ ray extent ⁇ [ T ⁇ ( ⁇ ⁇ + ⁇ ⁇ ⁇ i ⁇ ⁇ ) + S ⁇ ( ⁇ ⁇ + ⁇ ⁇ ⁇ i ⁇ ⁇ ) ] ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ⁇ ( 12 )
  • i ⁇ right arrow over ( ⁇ ) ⁇ represents a unit vector pointing in the direction of the angle vector ⁇ right arrow over ( ⁇ ) ⁇ .
  • ⁇ right arrow over ( ⁇ ) ⁇ is a simple angle in the plane in 2D, but becomes a two dimensional position on the unit sphere in 3D (in mathematical terms ⁇ right arrow over ( ⁇ ) ⁇ is a point in S n , the n dimensional sphere).
  • ⁇ right arrow over ( ⁇ ) ⁇ ⁇ right arrow over ( ⁇ ) ⁇
  • ⁇ right arrow over ( ⁇ ) ⁇ sin ⁇ cos ⁇ i x +sin ⁇ sin ⁇ i y +cos ⁇ i z .
  • the additive mapping is most useful in cases where the source is considered as a coherent field and the detection process is nonlinear, as is the case with intensity detectors.
  • ⁇ R is the field due to the reference structure and ⁇ S is the field due to the source. If the reference structure is designed to produce a scattered or radiated field of a known frequency or known basis, it is possible to filter on this frequency or basis to isolate ⁇ S from the measurement, essentially achieving self-referencing structural holography.
  • RST uses the reference structure to eliminate the need for special illumination and for scanning sequences.
  • reference structure modulated measurement of a source may be termed self-referencing.
  • the coherent (additive RST case) or incoherent (multiplicative RST case) field radiated by the source provides its own interferometric (additive case) and/or projective (multiplicative case) reference fields by structuring source radiation patterns. Because a source in a reference structure modulated measurement does not need to be scanned as a function of time, the source space might also be termed “self-revealing.”
  • heavy atom methods use diffuse arrays of point scatterers to generate a reference signal.
  • a discrete set of strong point scatterers is embedded in a target molecule.
  • R ⁇ ⁇ ( k ⁇ ) ⁇ strong scatterers ⁇ ⁇ j ⁇ ⁇ k ⁇ ⁇ d ⁇ i .
  • Additive reference structure tomography is a generalization of heavy atom methods to weaker and more distributed scatterers and to novel sensor modalities in optics, acoustics and other propagating field modalities.
  • additive RST might consist, for example of binding unknown target molecules to known reference structure molecules to enhance x-ray determination of molecular structure.
  • RST can achieve multi-dimensional source reconstruction without the spatial and temporal scanning associated with conventional tomography.
  • RST can sort and filter radiation to obtain all projections in a single time step.
  • the reference structure determines the spatial sampling rate and limits the resolution of the source reconstruction, but this limit is not necessarily worse than the restrictions imposed by conventional sampling.
  • RST is used to estimate sources on multi-dimensional spaces without scanning, to estimate sources on unusual bases, such as modal densities as a function of time and to estimate sources using fewer measurements than na ⁇ ve system design might suggest.
  • the dynamic range meaning the number of different discrete values that the source distribution in each cell can assume, is D p . If the state of each cell is independent, the number of different states this source can assume is D p N p . Now suppose that we wish to estimate this source using N m measurements of dynamic range D m . The number of measurement states exceeding the number of possible source states is a necessary but not sufficient condition for discrimination of each source state. For the source described above this requirement implies that N m logD m ⁇ N p logD p . Conventional tomographic and imaging systems satisfy this requirement by making at least on measurement per resolution cell or pixel.
  • Reference structures enable the receiver pattern 126 illustrated in FIG. 16 to be implemented with multiple spatial resolutions and to combine source points in coded patterns. Receiver pattern programming enables RST to improve sensor efficiency over conventional systems in cases where the state of different resolution cells are not independent. Few sources exhibit such independence in practice, in general, the state of a given cell is influenced by the states of adjacent cells. For example, in an image of a face, the presence of an eye in one region of the image implies that another eye will appear elsewhere in the image. The lack of independence of different regions of the image is the basis of image compression algorithms, which commonly reduce the raw image data by 1-3 orders of magnitude. Reference structure tomography can be used to implement such compression directly in the physical sensor, so that the amount of data sensed and processed is substantially reduced with little or no cost to system performance.
  • the number of sources may be limited in a variety of ways.
  • the source may consist of a discrete set of objects or patterns.
  • the source consists of one point radiator occupying one of N p resolution cells. This situation often occurs in motion imaging systems, in which at most one resolution cell is changing as a function of time. The problem is easily generalized to any finite number of occupied cells, however. A single occupied cell produces at most N p states.
  • reference structures to shape the spatial response pattern of, for example, a microphone for multi-dimensional acoustic imaging or the response of a pyroelectric detector for infrared tracking and identification.
  • ⁇ s j ⁇ is the discretized source density and ⁇ ij ⁇ is the discretized receiver pattern created by the reference structure.
  • the address of the occupied cell is the binary number ( m log 2 ⁇ N p m 1 ) ⁇ ( m log 2 ⁇ N p - 1 m 1 ) ⁇ ( m log 2 ⁇ N p - 2 m 1 ) ⁇ ⁇ ⁇ ( m 3 m 1 ) ⁇ ( m 2 m 1 ) ( 17 )
  • FIG. 17 illustrates a geometry for analysis of spatial scaling vs. resolution for two-dimensional RST.
  • the source 128 is described by a density function s(x,y).
  • computed tomography imaging is commonly implemented by digital inversion of a sampled set of ray projections of the density function through the object to be imaged. In conventional implementations, these projections are sampled using a temporal sequence of illumination and reconstruction geometries. These projections are measured and interpreted in the context of a source and illumination model.
  • l is the shortest distance from the origin to the ray and ⁇ is the angle the ray forms with respect to the x axis.
  • S is the Fourier transform (FT) of s, meaning that the FT of a projection is the slice of the FT of the source.
  • FT Fourier transform
  • samples in the Fourier space are polar sampled on a disk of radius 2 ⁇ / ⁇ l.
  • the maximum sampling period on the disk is the greater of 2 ⁇ /R s and 2 ⁇ / ⁇ l.
  • a reference structure can eliminates to scan in obtaining tomographic projections.
  • a radiation source lies within a circle 136 of radius R s , which lies within a measurement circle 130 of radius R m .
  • Rays 132 passing through the source space and incident on the measurement circle are also illustrated.
  • the rays are shown in bundles for values of l spaced by ⁇ l relative to each other, where l is defined as in FIG. 17.
  • the angular displacement from one bundle of rays to the next is ⁇ , where ⁇ is also defined as in FIG. 17.
  • the separation ⁇ l between the sampling rays corresponds to the pixel spacing on the measurement circle. Note that the sampling rays cross in the region adjacent to the source. We can insure that the only one ray is incident on each sensor point by setting tan ⁇ ( R s R m ) ⁇ ⁇ ⁇ ⁇ .
  • Bundles of holes spanning the R s radius of the inner circle would be formed as shown in the figure. These bundles would be spaced on angular displacements of ⁇ . Detectors at the ends of each hole would measure the transmittance or radiance of the source along the corresponding ray path, thus sampling g(l, ⁇ ) as in Eq. (18).
  • FIG. 19 illustrates that it is possible to build a 2D reference structure that can reconstruct a 2D source tomographically without scanning as long as the radius of the reference structure is greater the square of the radius of the source divided by the spatial resolution of the reconstruction.
  • This argument can be extended to higher dimensions by extending ray projections as in FIGS. 18 and 19.
  • FIG. 20 shows a reference structure analogous to FIG. 19 for a 3D source. The structure consists of layers of 2D reference structures 138 spaced by ⁇ z. As in FIG.
  • the black rays in the figure would be holes between detectors on the measurement cylinder.
  • the outer measurement cylinder 132 is of radius R m .
  • the inner cylinder 139 would be a hollow container for the source under measurement of radius R s Separation of the rays requires that the hole diameters be less than ⁇ z and ⁇ l and that R m >R s 2 / ⁇ l, as above.
  • i ⁇ right arrow over ( ⁇ ) ⁇ s a unit vector in the direction of the angles ⁇ right arrow over ( ⁇ ) ⁇ ( ⁇ , ⁇ ) on the unit sphere.
  • i ⁇ right arrow over ( ⁇ ) ⁇ 1 and i ⁇ right arrow over ( ⁇ ) ⁇ 2 are unit vectors in three dimensions orthogonal to i ⁇ right arrow over ( ⁇ ) ⁇ and to each other.
  • the geometry of the integration of Eq. (20) is illustrated in FIG. 21.
  • a detector at the measurement point 133 with an unobstructed view of the plane spanned by i ⁇ right arrow over ( ⁇ ) ⁇ 1 and i ⁇ right arrow over ( ⁇ ) ⁇ 2 a distance l away from the origin measures g(l, ⁇ right arrow over ( ⁇ ) ⁇ ) as in Eq. (20).
  • the cross section of the point of the cone at the measurement point determines the reconstruction resolution ⁇ x. If one uses a sampling scheme that full populates the outer sphere with sampling points, equality between the number of data points on the measurement sphere and the number of reconstructed source points will require R m 2 >R s 3 /3 ⁇ x. Many sampling schemes may approach this limit.
  • ⁇ ij and m ⁇ are angular positions on the measurement sphere. This strategy produces approximately 4 ⁇ R m 2 / ⁇ x samples.
  • Radon transform-based reference structures illustrate that reference structures can be constructed to enable scan-free multi-dimensional sensing, the light efficiency of these structures is relatively poor. In the ray based structures the power detected by the sensors is only 1 N p
  • N p is the number of resolution elements in the source.
  • the 3D planar sampling reference structures obtain somewhat better performance, detecting approximately N p - 1 3
  • Reference structures can be designed to provide multiplex coverage with a larger system response function, which improves photon efficiency. Radon based transforms described designed such that only ⁇ square root ⁇ square root over (N p ) ⁇ source cells are detected by each detector in the 2D case and N p 2 3
  • [0155] are detected in the 3D case.
  • signals from all the source cells are radiated to each detector element and unwanted signals must be obscured or absorbed to enable Radon-based coding.
  • Hadamard and Fourier-transform based schemes might be implemented on the source space to improve photon efficiency.
  • Hadamard and Fourier systems measure approximately half of the power incident on each detector element. We describe these approaches further below when considering detailed implementations.
  • a point source as viewed through the masks, can be analyzed as follows.
  • the transmittance of the i th mask is t i (x), as shown in the schematic representation of FIG. 23 b , where the source 162 projects a signal ray 168 through a series of masks 164 to a detector 166 .
  • R is the range to the source.
  • m ⁇ ( x ) ⁇ i ⁇ ( 1 + cos ⁇ ( K ⁇ ⁇ x ⁇ ( 1 + i ⁇ ⁇ ⁇ R ) - K ⁇ i ⁇ ⁇ ⁇ ⁇ ⁇ x s R ) ) ( 25 )
  • FIG. 25 is a transmission graph of the same system with the number of masks increased from 10 to 100.
  • FIG. 26 is a transmission graph utilizing the same parameters used to generate the graph of FIG. 25, but where the cosine function has been replaced with a rectilinear function.
  • FIG. 27 is a transmission graph of an RST system with x fixed at 0 and a function of x S for a fixed x.
  • FIG. 30 is a schematic diagram of a plane wave motion sensor model, including a motion sensor as a mapping from plane waves onto measurements.
  • a plane wave incidence signal 200 is received by a reference structure 204 at an angle of ⁇ to the normal 202 to the surface of the reference structure 204 .
  • Sensor points 206 are disposed behind the reference structure 204 and configured to receive any signal from the reference structure 204 .
  • the reference structure 204 consists of N layers of transmittance masks and that the incident signal 200 is f( ⁇ ).
  • FIGS. 32 and 33 are graphs that show the 3D output of such a configuration by sampling longitudinally with masks.
  • FIG. 34 is a schematic diagram of mask and sensor geometry and shows a transparency 232 and a sensor assembly 236 , where the sensor assembly 236 includes and 8 ⁇ 8 array of sensors 238 of D ⁇ D dimensions. Ray projections 234 are received by sensors 238 from the transparency 232 . As illustrated in FIG. 34, the distance from the sensor 236 plane to the entrance aperture of the transparency 232 must be sufficient to separate rays 234 onto adjacent channels or sensors 238 , i.e. L> ⁇ / ⁇ .
  • FIG. 35 is a grid showing how the measurement pixels are numbered in the sensor assembly 236 of FIG. 34. The grid position as a function of pixel number for an n ⁇ n array can be determined, for example, with the computer program listed in Appendix A.
  • Connectivity can be assigned using the algorithm listed in Appendix A.
  • Different views of the connectivity pattern of the reference structure for the reduced Hadamard connectivity are illustrated in FIGS. 36 a - 36 c at various angles of view where each connectivity pathway 242 , runs between a reference structure entrance 240 and a reference structure exit 244 .
  • FIG. 37 is a graph which shows the singular value spectrum associated with this interconnection pattern.
  • FIG. 38 is a schematic diagram of reference structure pipes 246 with a length of L, a width of d and an aperture angle ⁇ .
  • a ray 250 may travel through the pipe 246 off-axis up to an angle ⁇ .
  • there are 100 micron pixels configured to detect over a ⁇ 10 degree range, meaning that each ray covers 2.5 degrees.
  • FIG. 39 is the reduced S matrix.
  • FIG. 40 is the connectivity pattern for the configuration shown in FIG. 38. In FIG. 40, the connectivity paths 254 lie between the entrance 252 to the reference structure and the exit 256 to the reference structure.
  • FIGS. 41 a - 41 k are cross-sectional schematic views of a set of masks, in order from an input side to an output side where each mask 258 a - 258 k has multiple apertures 260 . Each mask 258 a - 258 k may be stacked or layered to form the corresponding reference structure. As shown in FIGS. 41 b - 41 i , the apertures 260 of masks 258 b - 258 i may overlap other apertures 260 .
  • the number of angular resolution elements along each transverse dimension is N ⁇ A/ ⁇ , where A is the full transverse aperture.
  • the transverse angular range observed is ⁇ .
  • L ⁇ 2 A/ ⁇ is obtained for the thickness of the reference structure for a given aperture size and angular range.
  • FIG. 42 is a schematic of a system for imaging a 4 ⁇ 1 LED array using a reference structure.
  • the reference structure 263 consisting of a solid opaque block with transparent paths or pipes 264 therethrough.
  • the transparent pipes 264 run from the input aperture 262 of the reference structure 263 to the exit aperture 266 .
  • Such a reference structure 263 can be made by fabrication techniques including selective laser sintering, stereolitograph and holography.
  • the reference structure 263 is designed to accept plane waves at the reference structure input aperture 262 at 4 angles of approximately ⁇ 10°, ⁇ 5°, 5°, 10°.
  • the reference structure 263 consists of 10 of the transparent pipes 264 of 1 mm diameter, each oriented at the aforementioned angles at the input 267 of the reference structure 263 .
  • Each of these pipes 264 goes to one of the four sensors in a sensor array space close to the reference structure 263 .
  • Each sensor may receive a combination of these pipes 264 as shown in FIG. 42.
  • a sensor array 272 receives light from LEDs 267 of an LED array 268 through a reference structure 270 .
  • the LED array 268 was used as the source to characterize the matrix representing the reference structure 270 .
  • the LED array 268 consisted of 4 LEDs 267 spaced 5 cm apart, and was placed 50 cm from the reference structure 270 .
  • a 12 bit cooled CCD camera was used as a sensor array 272 .
  • Each sensor was chosen to be a group of 60 ⁇ 50 pixels (pixel size 4.65 ⁇ 4.65) on the cameras imaging surface.
  • FIGS. 44 a - 44 o show the patterns 274 a - 274 o transmitted to the camera by the reference structure 270 for different combinations of lighted LEDs 267 .
  • the intensity values corresponding to the LED pattern 1000, 0100, 0010, 0001 constitute the columns of the matrix, which characterize the reference structure 270 represented by the matrix.
  • H [ 0 538590 1197105 0 0 733136 0 881874 301831 623103 1743120 330926 487603 257955 0 0 ]
  • Infrared motion sensors are widely used in security and control applications and can be improved with the addition of a reference structure.
  • Conventional motion sensors consist of two or more detectors operating in differential mode, where the measurement of each detector is referenced against the other. As a source moves through the receiver pattern of these detectors, changes in the states of the detectors relative to one another occur.
  • the receiver pattern of such a detector is a 3D density map of the probability that a photon emitted at a point in space hits the detector.
  • FIG. 45 a schematic of an example of a receiver pattern for a conventional 2-element motion sensor is shown.
  • two detectors 276 , with optics 277 exhibit a receiver pattern with two detection areas 278 and 280 .
  • FIG. 46 is a schematic in which the two detectors 276 with optics 277 have a receiver pattern with three detection areas 282 interposed between 3 non-detection areas 284 .
  • An embodiment of the invention expands the utility of motion sensors to include object recognition and tracking capabilities by making receiver patterns more complex.
  • Receiver patterns of conventional motion detectors are shaped to discriminate, for example, between people and animals or vehicles. These patterns are convex, however, meaning that the receiver pattern is a beam, shaped mostly like the target.
  • the receiver patterns are structured such that the temporal signature on the detectors as the source moves through the field of view corresponds to the range to the source. Since the receiver patterns will vary with wavelength of the radiation emitted by or which otherwise propagates from the source, the temporal pattern may also give a spectral signature for the source, allowing for source identification as well as tracking. Complex receiver patterns are particularly attractive as embedded processors and wireless communications ports are added to motion sensors.
  • FIG. 47 is an example of a receiver pattern on a wireless array.
  • two detector arrays 288 and 290 with corresponding optics 289 and 291 provide a receiver pattern forming a spatio-spectral grid 286 with intersecting detection areas 285 to detect the presence and motion of a source 292 .
  • the spatio-spectral grid 286 which the source 292 moves through on the array, will produce a temporal signal on the array of detectors 288 and 291 which allows for the determination of a characteristic of the source, and its position and velocity.
  • An embodiment of the invention in which the motion sensor is enclosed within a reference structure may be used as a densely deployed and integrated component of a heterogeneous environment of sensors such as acoustic motion sensors, seismic sensors, video and IR cameras, etc.
  • reference structures include 3D and 2D non-monotonic structures.
  • example structures may have transmittance functions based on Hadamard transforms (i.e., Hadamard matrices).
  • reference structures may be implemented using materials that modulate absorption, permittivity, impedance, florescence or other field propagation properties.
  • RST phase sensitive crystallography
  • RST can be used to sense the phase of multi-dimensional scattered fields. This sensitivity may be useful in macromolecule structure analysis. It may be preferable to implement RST by binding a well-characterized reference molecule to an unknown source molecule in a crystal structure. The reference scattering enables the phase sensitive Fourier representation of the unknown molecule to be determined.
  • FIG. 48 is a perspective ghost view of a reference structure 154 , in accordance with one embodiment of the present invention.
  • the reference structure 154 has an input side 160 and an output side 150 .
  • the input side 160 includes multiple input windows 158 .
  • the output side includes multiple output windows 152 .
  • each output window 152 is connected to multiple input windows 158 via transparent pathways 156 through the reference structure 154 , while the remainder of the reference structure 154 body is opaque.
  • a prototype version of reference structure 154 has been constructed using a stereolithographic printer. A photograph of the sensor is shown in FIG. 49.
  • FIGS. 50 and 51 show the reference structure signal as a function of time and channel number
  • FIG. 51 shows the inverse transform for source position (a car going by) as a function of time.
  • N orderOfGrid ⁇ ( NbyNgrid )
  • p 1 ⁇ [ N / 2 , N / 2 ] Neven [ ( N + 1 ) / 2 , ( N + 1 ) / 2 ]
  • Nodd p 2 p 1 + [ 0 , 1 ]
  • p 3 p 1 + [ 1 , 1 ]
  • p 4 p 1 + [ 1 , 0 ]
  • p EvenSquares p 1 + [ N / 2 , 1 - N / 2 ]
  • OddSquares p 1 - [ ( N - 1 ) / 2 , ( N - 1 ) / 2 ]

Abstract

A reference structure tomography device is provided which includes a reference structure configured to intercept and modulate energy in the form of waves or otherwise propagating from a source to a sensor, along longitudinal and traverse directions. The reference structure modulates or otherwise conditions the propagating wave to simplify an inversion process on the data set created by the interaction between the wave and the sensors. The reference structure can modulate a wave through multiple types of interactions with the wave including obscuring, defracting, defusing, scattering, and otherwise altering any characteristic of a portion of the wave. By selecting a reference structure that is compatible with the sensors, the number of measurements needed to resolve the source through the source wave is reduced. The reference structure can also increase the resolution of an imaging system. Thus, by reducing or altering the data collected by the sensors, the reference structure tomography device can improve the imaging abilities of the system.

Description

    CROSS REFERENCE TO RELATED APPLICATIONS
  • This application claims priority to Provisional Patent Application No. 60/414,362, filed Sep. 30, 2002 and Provisional Patent Application No. 60/414,344, filed Sep. 30, 2002. The contents of both of the above-identified applications are incorporated here in by reference in their entirety.[0001]
  • BACKGROUND OF THE INVENTION
  • 1. Field of the Invention [0002]
  • The present invention relates generally to sensors and imaging systems, and more particularly to tomographic systems. [0003]
  • 2. Background of the Related Art [0004]
  • Tomography involves the measurement of a set of source projections from a source, and inverting the measurement to estimate source characteristics or variables. The projections typically consist of electromagnetic or acoustic energy, but can also include any form of energy, which can be caused to propagate through a medium. The most common method of tomographic imaging inverts a discrete set of projection measurements to estimate source distribution as a point wise projection in normal or hyperspectral space. The inverted set of projections can yield an image of the interior of the medium through which the projections passed. Tomography finds application in a wide range of detecting and imaging fields, such as medicine, seismology, and non-destructive testing. Tomography is a calculation intensive process done numerically by computers, and thus the process is also referred to as “computer tomography.”[0005]
  • In most tomographic imaging systems, image reconstruction requires long computation times, due to the large number of data points and processing steps required. In addition, achieving multi-dimensional source reconstruction with traditional tomography requires the use of spatial and temporal scanning. [0006]
  • SUMMARY OF THE INVENTION
  • An object of the invention is to solve at least the above problems and/or disadvantages and to provide at least the advantages described hereinafter. [0007]
  • Therefore, the object of the present invention is to provide a reference structure tomography device and method that includes a multi-dimensional reference structure that is adapted to modulate electromagnetic fields along both longitudinal and transverse directions. [0008]
  • To achieve at least the above objects, in whole or in part, provided is a source revolver, comprising a multi-dimensional reference structure configured to modulate source radiation in a pre-determined manner to produce a reference structure modulated signal comprising a puerility of source space projections, where in the multi-dimensional reference structure comprises a longitudinal dimension, and an analyzer configured in accordance with said reference structure for resolving a source state from the reference structure modulated signal. [0009]
  • To achieve at least the above objects, in whole or in part, there is further provided a source imager, comprising a multi-dimensional reference structure configured to modulate source radiation to produce a reference structure modulated signal comprising a puerility of source space projections where in the multi-dimensional reference structure comprises a longitudinal dimension, a sensor array configured to detect the reference structure modulated signal, and a processor coupled to the sensor array and configured to derive source information from the detected reference structure modulated signal. [0010]
  • To achieve at least the above objects, in whole or in part, there is further provided a tomographic apparatus, comprising a source, a sensor, a multi-dimensional reference structure between the source and sensor for producing modulated radiation from source radiation, where in the modulated radiation is received by the sensor and comprises a puerility of a source space projections and where in the multi-dimensional reference structure comprises a longitudinal dimension, and an analyzer coupled to the sensor for analyzing the modulated radiation. [0011]
  • To achieve at least the above objects, in whole or in part, there is further provided a method of enhancing a sensor, comprising altering radiation received by the sensory by inserting a multi-dimensional reference structure into a path of the radiation, where in the altered radiation comprises a puerility of source space projections, and where in the multi-dimensional reference structure comprises a longitudinal dimension, and analyzing the sensors response to the altered radiation. [0012]
  • To achieve at least the above objects, in whole or in part, there is further provided a sensory system, comprising a multi-dimensional reference structure configured to receive unconditioned data radiated from a source and to produce condition data comprising a puerility of source space projections, where in the multi-dimensional reference structure comprises a longitudinal dimension, a sensory configured to receive the conditioned data, and an analyzer configured to invert the conditioned data received by the sensor. [0013]
  • To achieve at least the above objects, in whole or in part, there is further provided a measurement system, comprising a source configured to produce source modulated radiation, a multi-dimensional reference structure comprising a longitudinal dimension for receiving the source modulated radiation and producing reference modulated wave radiation comprising a puerility of source space projections, and an analyzer in communication with the reference modulated radiation for analyzing the reference modulated radiation. [0014]
  • To achieve at least the above objects, in whole or in part, there is further provided a method of resolving a source state, comprising modulating source radiation such that a puerility of source space projections are created in a parallel fashion, and resolving the source state using a puerility of source space projections. [0015]
  • Additional advantages, objects, and features of the invention will be set forth in part in the description which follows and in part will become apparent to those having ordinary skill in the art upon examination of the following or may be learned from practice of the invention. The objects and advantages of the invention may be realized and attained as particularly pointed out in the appended claims. [0016]
  • BRIEF DESCRIPTION OF THE DRAWINGS
  • The invention will be described in detail with reference to the following drawings in which like reference numerals refer to like elements wherein: [0017]
  • FIG. 1 is a schematic representation of a reference structure tomography (RST) system, in accordance with one embodiment of the present invention; [0018]
  • FIG. 2 is a schematic representation of a RST system, in accordance with another embodiment of the present invention; [0019]
  • FIG. 3 is a schematic representation of a RST system, in accordance with another embodiment of the present invention; [0020]
  • FIGS. [0021] 4-7 are schematic diagrams of various RST devices, in accordance with the present invention;
  • FIG. 8 is a schematic illustration of a basic structure for RST imaging, in accordance with the present invention; [0022]
  • FIG. 9 illustrates the ray geometry of FIG. 8; [0023]
  • FIG. 10 illustrates a point source geometry; [0024]
  • FIGS. [0025] 11-15 are schematic diagrams illustrating a reference structure obscuring various point sources, in accordance with the present invention;
  • FIG. 16 is a schematic diagram of a RST system, in accordance with another embodiment of the present invention; [0026]
  • FIG. 17 illustrates a geometry for analysis of spatial scaling vs. resolution for two-dimensional RST, in accordance with the present invention; [0027]
  • FIG. 18 illustrates a measurement circle; [0028]
  • FIGS. [0029] 19-20 are three-dimensional schematic representations of RST systems, in accordance with the present invention;
  • FIG. 21 is a schematic illustration of a RST system that utilizes spherical geometry and obscuring sphere, in accordance with one embodiment of the present invention; [0030]
  • FIG. 22 is a schematic illustration of a reference structure formed around clear polygons in accordance with the present invention; [0031]
  • FIG. 23[0032] a shows channels for Δx=Rs/100, in accordance with the present invention;
  • FIG. 23[0033] b is a schematic representation of stacked reference structure, in accordance with the present invention;
  • FIGS. [0034] 24-29 are transmission graphs of the reference structure of FIG. 23;
  • FIG. 30 is a schematic diagram of a plane wave motion sensor, in accordance with the present invention; [0035]
  • FIG. 31 illustrates a connection matrix for the plane wave motion sensor of FIG. 30; [0036]
  • FIGS. [0037] 32-33 are graphs that show the 3-dimensional output of the connection matrix of FIG. 31, in accordance with one embodiment of the present invention;
  • FIG. 34 is a schematic diagram of mask and sensor geometry, in accordance with one embodiment of the present invention; [0038]
  • FIG. 35 is a grid showing how measurement pixels may be numbered in the mask and sensor geometry of FIG. 34; [0039]
  • FIGS. 36[0040] a-36 c illustrate different views of a connectivity pattern for a reference structure, in accordance with the present invention;
  • FIG. 37 is a graph of the singular value spectrum associated with the connectivity patterns of FIGS. 36[0041] a-36 c;
  • FIG. 38 illustrates a reference structure pipe with a length of 4.5 mm, in accordance with one embodiment of the present invention; [0042]
  • FIG. 39 illustrates a reduced S matrix for the configuration of FIG. 38; [0043]
  • FIG. 40 illustrates a connectivity pattern for the reference structure of FIG. 38 [0044]
  • FIGS. 41[0045] a-41 k are cross-sectional schematic views of a set of masks, in order from an input side to an output side, which may be stacked or layered to form a reference structure in accordance with an embodiment of the present invention;
  • FIGS. [0046] 42-43 are schematic diagrams that illustrate imaging of a 4×1 LED array using a reference structure, in accordance with an embodiment of the present invention
  • FIGS. 44[0047] a-44 o illustrate reference structure output patterns, in accordance with an embodiment of the present invention;
  • FIG. 45 illustrates a receiver pattern for a conventional motion sensor; [0048]
  • FIG. 46 illustrates a receiver pattern for a motion sensor with a reference structure, in accordance with an embodiment of the present invention; [0049]
  • FIG. 47 illustrates a receiver pattern for a motion detector with multiple reference structures, in accordance with an embodiment of the present invention; [0050]
  • FIG. 48 is a perspective ghost view of a reference structure, in accordance with one embodiment of the present invention; [0051]
  • FIG. 49 is a photograph of a reference structure, in accordance with one embodiment of the present invention; and [0052]
  • FIGS. [0053] 50-51 are photographs of output signals for the reference structure of FIG. 49.
  • DETAILED DESCRIPTION OF PREFERRED EMBODIMENTS
  • The present invention utilizes the modulation of a selected physical property or set of properties of energy propagating between a source and a receiver in such a way that the receiver produces an enhanced measurement. The utilization of this technique is hereinafter referred to as “Reference Structure Tomography” (RST). [0054]
  • As used herein, a “source” can be either a generator, scatterer or absorber of propagating energy, and may be described by a source distribution function. Accordingly, a source can either emit energy, or can receive and transmit, obscure, reflect, refract, diffract or otherwise interact with propagating energy. The propagating energy or projection is typically modeled as following a ray path through or from the source. A source projection is a sum or integral of properties associated with a subset of the points in the source. The inversion of source projections can be done through numerical methods executed by a computer. [0055]
  • RST uses reference modulations in the space between sources and sensors to encode propagating energy, also referred to as fields, radiated or scattered from the sources. The reference modulations are preferably achieved by using a reference structure. [0056]
  • RST preferably utilizes multi-dimensional reference structures to precondition projections for detection and/or subsequent inversion. Various RST embodiments will be described below. For example, “embedded detector RST” is an embodiment in which the detector elements are contained within the reference structure. “Adaptive RST” is an embodiment in which the reference structure changes or is changed in response to the state of the source or the measurement. “Impulse RST” is an embodiment in which the projection is similar to an impulse response function, and the impulse response function of a RST system is the measurement produced by an impulse in the source space as a function of the position of the impulse. [0057]
  • As used herein, the term “projection” is synonymous with the terms “radiating field,” “ray,” “wave,” “radiation,” and “propagating energy,” etc. A reference structure may consist of a multi-dimensional distribution structure both transverse and longitudinal to primary directions of propagation for the projection. If the radiation space is two-dimensional, such as waves propagating along a surface, the reference structure may also be two-dimensional. If the radiation space is three-dimensional such as waves propagating through a volume, the reference structure may also be three-dimensional. [0058]
  • A reference structure can be thought of as conditioning or encoding the radiation. The goal of the encoding process is to condition source characteristics for computational estimation from sensor measurements. The most typical objective is the estimation of the source distribution over a multi-dimensional space from measurements over a lower dimensional space. For example, one may seek to estimate the source density in a plane from measurements on a curve or one may seek to estimate the source density in a volume from measurements over a surface. The simplest form of RST involves the estimation of sources from geometric projections modulated by obscurations. [0059]
  • Referring to FIG. 1, a schematic representation of a RST system is shown. In the schematic representation, a [0060] linear transformation space 14 receives source data 12 from a source space 10. The linear transformation space 14 transforms the source data 12. The transformed source data 16 propagates from the linear transformation space 14 and is received by a sensor space 18. This diagram is a generalized representation of a RST system, whereby the linear transformation space 14 may correspond to a reference structure, the source space 10 can correspond to a source, and the sensor space 18 can correspond to a sensor or sensor array. The source data 12 and the transformed source data 16 correspond to propagating energy, such as electromagnetic radiation or acoustic waves, before and after interaction with the linear transformation space 14, respectively.
  • FIGS. 2 and 3 are schematic representations of two examples of a RST system. In FIG. 2, a [0061] source 20 is surrounded by a reference structure 21, which consists of obstructions 22, and the reference structure 21 is surrounded by sensors 24. The circular geometry of FIG. 2 shows how a source 20 can be completely surrounded by the reference structure 21 and the sensors 24, suitable for a circular coordinate system.
  • In FIG. 3, a radiating [0062] field 26 is received by a reference structure 28, which is placed between the radiating field 26 and the detectors 30. The reference structure 28 consists of obstructions 22. The geometry of FIG. 3 has a transverse direction and longitudinal direction, and shows how a reference structure and detectors can be modeled in a rectilinear coordinate system.
  • In the above examples, a direct projection is a bundle of ray projections integrated across a range of angles. A ray projection may be considered a visibility projection integrated as the angular range approaches zero. A diffuse projection is a projection for points not distributed along a line, and a diffractive projection is a diffuse projection generated by diffractive wave propagation. [0063]
  • RST offers the advantage of extended depth of field and other data cube projections. In general, the term data cube originates in hyperspectral imaging, and means a two dimensional image viewed as a function wavelength or frequency forms a “data cube” consisting of a stack of images at different colors. Images in higher spaces, such as a three dimensional source viewed as a function of color, may form multi-dimensional or hyper data cubes. Generalizing to a higher level, one may view the data cube as the set of all spatially related measures of the state of a source. As used herein, the data cube refers to the cube describing the field intensity of a three dimensional source. Extended depth of field consists of creating a 2D image from a 3D source by integrating line integrals through the source (a 3D source consists of a stack of images. An extended depth of field image smashes all of these images into a single plane. [0064]
  • Extended depth of field or slices of the data cube as a function of space or color are conventional projections of the data cube. A projection is an integral of the form m=∫h(r)S(r)dr, where S(r) is a density function describing the source at data cube position r and h(r) is a projection vector. For extended depth of field, h would be a delta function along a line. For monochromatic imaging, h would pick a color of the data cube. An unconventional projection would be something other than the projection that picked a slice or a line of the data cube. [0065]
  • As used herein, “diffractive field propagation” is propagation of the field according to a wave equation, i.e., normal propagation of an optical or acoustic field, and a “spatio-spectral source” is a source described by a data cube that includes both spatial and color axes. [0066]
  • The term “basis” is used herein to mean a set of vectors or functions that describe a source. For example, a source may be described by the amplitudes of impulse response functions sampled on a regular spatial grid. Alternatively, a source may be described by a Fourier series, in which case the basis consists of sine and cosine functions. A “non-standard basis” would be a description of a source using basis functions that are not regular sampling functions on a Euclidean space or that are not a simple Fourier basis. [0067]
  • Even if a RST mapping does not immediately enable multi-dimensional reconstruction, RST systems may map the source space onto the measurement space so as to immediately produce unconventional projections of the source space. A measurement of the visible ray projection of the source space, even in systems dominated by diffractive field propagation, is an example of such a system. Other examples include systems that project the spatio-spectral source data cube onto nonstandard basis. [0068]
  • If one seeks to reconstruct an arbitrary source distribution S(r), RST enables reconstruction without scanning or with reduced scanning. Related art computer tomography scans an illumination and/or detector system to serially acquire projections. A reference structure sorts and creates projections such that complete data can be obtained in a single time step. The number of measurements made for a given source size and resolution is approximately the same in both cases. Another advantage of RST arises when the source can be represented by a set of basis functions with fewer components than the related art locally sampled basis. For example, if only one or a few source points are active, the source can be represented by a number of basis functions equal to the logarithm of the number of spatial resolution cells. In such cases, RST can reconstruct the source using many fewer measurements than conventional systems. RST can be more efficient because it is a multiplex technique. Multiplexing enables one to sort and find sources in a way that is more efficient than measuring all possible source states. RST also offers the possibility of revealing more information about a measured space by increasing the complexity of the energy received by the sensors. [0069]
  • Multiplex sensing consists of measurements that depend jointly on multiple source variables. A projection which depends on the state of a source at multiple points is a multiplex measurement. Tomography relies on multiplex sensing and can be implemented using detector elements external to a reference structure that is integrated between the source and detector elements, or by using detector elements wholly or partially embedded in the reference structure. RST may be implemented with static or time varying reference structures. [0070]
  • FIG. 4 is a schematic diagram of another embodiment of a [0071] RST system 100. The RST system of FIG. 4 includes a source 32 from which a source wave 34 emanates. A reference structure 36 receives the source wave 34, and a reference structure modulated wave 38 propagates from the reference structure 36. The reference structure modulated wave 38 is received by a sensor 40. The sensor 40 produces a signal in accordance with the reference structure modulated wave 38 and directs that signal to a processor 44 through a signal line 42 between the sensor 40 and the processor 44.
  • In operation, the [0072] RST system 100 is configured so that the reference structure 36 receives a source wave 34 from the source 32, and provides a reference structure modulated wave 38 which is received by the sensor 40. The reference structure 36 modifies the source wave 34 in accordance with an interaction between the source wave 34 and the reference structure 36. Thus, the reference structure modulated wave 38 is modulated based on a combination of the characteristics of the source 32 from which the source wave 34 originated and the reference structure 36. The processor 44 is configured to invert the response of the sensor 40 using a predetermined transformation function of the reference structure 36 to characterize the source 32.
  • FIG. 5 is a schematic diagram of a [0073] RST system 200 with an active reference structure 54, in accordance with another embodiment of the present invention. The active reference structure 54 receives the source wave 52 and a reference structure modulated wave 56 propagates from the active reference structure 54. The active reference structure 54 also receives a signal from a reference structure driver 68 through a signal line 70 between the active reference structure 54 and the reference structure driver 68. The reference structure modulated wave 56 is received by an analyzer 58. The analyzer 58 includes a sensor 60, which is connected to a processor 64 through a signal line 62. The analyzer 58 is coupled via a signal line 66 to the reference structure driver 68. In operation, the active reference structure 54 receives a feedback signal from the analyzer 58 and the reference structure driver 68. The active reference structure 54 can alter its characteristics in accordance with variations in the source wave 52 as detected by the sensor 60 from the reference structure modulated wave 56.
  • The [0074] RST system 200 also includes a wave generator 46, which generates an illumination wave 48. The illumination wave 48 is received by the source 50 and the source wave 52 then propagates from the source 50. Accordingly, a source wave 52 is not necessarily originally generated by the source 50, but can be instead a wave which can either be generated by the source 50, or can be received and reflected, transmitted, or otherwise propagated from the source 50.
  • FIG. 6 is a schematic diagram of a [0075] RST system 300, in accordance with another embodiment of the present invention, utilizing the wave generator 46, illumination wave 48, source 50 and source wave 52 of FIG. 6. In this embodiment, a point type reference structure 72 is used from which a reference structure modulated wave 56 propagates. The reference structure modulated wave 56 is received by a plurality of detectors 60, which form a sensor array 74 with sampling points corresponding to each of the detectors 60. The sensor array 74 has a signal line 76 coupled to a processor 64. In operation, the discrete sampling points (detectors 60) of the sensor array 74 of the RST system allow the reference structure modulated wave 56 to be discretely sampled.
  • FIG. 7 is a schematic diagram of a [0076] RST system 400 with a source wave modulator, in accordance with another embodiment of the present invention. In addition to the wave generator 46, and illumination wave 48, this embodiment includes a source wave modulator 88 from which a modulated illumination wave 82 emanates to be received by the source 50. A source wave 84 from the modulated illumination wave 82 propagates from the source 50 and is received by the reference structure 36. A reference structure modulated wave 86 then propagates from the reference structure 36 to be received by a sensor 40. The source wave modulator 88 is modulated by a modulator driver 80 through a signal line 78. Accordingly, the source wave modulator 88 can modulate the illumination wave 48, so that the source 50 receives a modulated illumination wave.
  • Multiplicative Visible Projection RST [0077]
  • In this context, “visible” means detectable by the sensor and “visible projection” means that each sensor detects along the lines that are not obscured to it from the source. As discussed above, computer tomography consists of the measurement and numerical inversion of ray projections, and can be enhanced by the inclusion of a reference structure. A RST system may measure ray projections, but these projections will depend jointly on the source and the reference structure. Multiplicative reference structures are structures which either completely obscure or completely transmit the energy of the ray projection. [0078]
  • For example, FIG. 8 is a schematic illustration of the basic structure for a RST imaging system, for the case of ray projections. In FIG. 8, a [0079] source 20 is surrounded by a reference structure 90, and the reference structure 90 is surrounded by detectors 24. In operation, the detectors 24 receive energy from the source 20, as represented by ray projections 92, 94, 96.
  • For the geometry shown in FIG. 8, Assuming the energy travels as represented by the ray projections, the state of the i[0080] th detector is m i = detector area detector acceptance angle ray extent T i ( θ , ρ ) S ( ρ + α i θ ) α θ ρ , ( 1 )
    Figure US20040095626A1-20040520-M00001
  • where T[0081] i({right arrow over (θ)},{right arrow over (ρ)}) is the transmittance of the reference structure for rays incident at angle {right arrow over (θ)} relative to a ray normal to the detector, {right arrow over (ρ)} is a transverse position on the detector surface, and α is a distance along the ray in direction i{right arrow over (θ)}. Note that Ti({right arrow over (θ)},{right arrow over (ρ)}) is the x-ray transform of the reference structure (the Radon transform in 2D).
  • S(ρ) is the state of the source described as a function of position in space (i.e. a source density function.) In conventional tomography, S would be the source, e.g., if you look at a computer tomographic image of bone and tissue, S is the 3D function describing the intensity of the image as a function of position. α is a scalar parameter of integration. An x-ray projection is the integral of S along a line through the source. In 2D, the set of all line integrals of S is the Radon transform. It is not quite the Radon transform in 3 and higher dimensions, but is related. The Radon transform is invertible for S, which forms the basis for 2D computer tomography. Computer tomography often reconstructs sources in 2D slices using Radon transforms. One obtains these transformations in computer tomography by scanning the illumination to get projections of the source along lines. Reference structure tomography (RST) uses the reference structure to get discrete groups of line or other projections without scanning, and “integration along the ray extent” means integration from one end of the ray to the other. The ray extent would be the length of the ray or the space spanned by it, and the state of the ith detector is the amplitude of the signal it is currently receiving. [0082]
  • FIG. 9 illustrates the ray geometry of FIG. 8. In FIG. 9, a [0083] detector 106 has dimensions P1×P2 and receives ray projections 98, 100, and 102. The normal vector to the surface of the detector 106 is 104. The angle between the normal vector 104 and ray projection 98 is θ2, and the angle between the normal vector 104 and ray projection 102 is θ1.
  • In the example shown in FIG. 9, the need for a reference structure is not immediately apparent because m[0084] i contains sufficient information to reconstruct the source space even if T({right arrow over (θ)},{right arrow over (ρ)})=δ({right arrow over (θ)}−{right arrow over (θ)}p)δ({right arrow over (ρ)}−{right arrow over (ρ)}p), i.e., if each measurement corresponds to a particular ray. However, the reference structure may reduce the number of time steps and number of detectors needed to characterize the source for purposes of numerical inversion. Note that the measurements mi constitute a discrete linear transformation of the source state and that this transformation is invertible.
  • An embodiment of visible projection RST may include a source consisting of a 2×2 array of point sources of light. The geometry of such point sources is illustrated in FIG. 10, and are labeled S[0085] 11, S12, S21, S22, corresponding to their position in an array. The sources may be measured, for example, by a detector array/reference structure combination, such as the one illustrated in FIGS. 11-15. In FIGS. 11-15, a reference structure containing obstructions 111 lies between the point sources S11, S12, S21, S22 and detectors 107, 108, 109 and 110.
  • As illustrated in FIG. 12, the reference structure of FIG. 11 obscures all point sources from the [0086] first detector 107 except for point source S12 because it lies outside the obscured area 112. FIG. 13 shows how the second detector 108 can detect only source points S11 and S21 because they lie outside the obscured area 114. FIG. 14 shows how the third detector 109 can detect only points S12 and S22 because they lie outside the obscured area 116. Finally, FIG. 15 shows how the fourth detector 110 can detect S12, S21 and S22, but not S11, because they lie outside the obscured area 118.
  • For a chosen geometry of detectors, reference structure and point sources, it is possible to model the detector and reference structure system's response to a source or sources with a matrix. For the system illustrated in FIGS. [0087] 11-15, the source-to-measurement transformation may be represented in matrix form as m=Ts where T = [ 0 1 0 0 1 0 1 0 0 1 0 1 0 1 1 1 ] ( 2 )
    Figure US20040095626A1-20040520-M00002
  • If the reference structure is designed such that T is nonsingular, than this transformation may be inverted to estimate the source from the measurements. The process of inversion consists of multiplying the measurement vector by the inversion matrix. This is called “algebraic inversion” and is always possible. In practice, more efficient inversion algorithms may be available and preferred. In the example of FIGS. [0088] 11-15 the inversion T−1 of T is T - 1 = [ 0 1 1 - 1 1 0 0 0 0 0 - 1 1 - 1 0 1 0 ] ( 3 )
    Figure US20040095626A1-20040520-M00003
  • Where the matrix T[0089] −1 is the linear transformation, it is preferable to design the reference structure such that the linear transformation from the source descriptors to the measurement is well conditioned for the estimation of desired source parameters.
  • A “source descriptor” describes the actual state of the source, however, a measurement is what is actually measured. A linear inversion or some other mathematical algorithm is used to estimate source descriptors or functions of source descriptors from the measurement. These estimates are what the source calls “source parameters,” also known as “source estimates” or “estimation parameters”. [0090]
  • As an example to illustrate the difference between source description and source parameters, the source descriptor could be “the state of the source is heads” or the “state of the source is tails”. The measurement could be a picture of the coin. Based on the measurement, one could estimate whether or not the state of the source is heads or tails. If the image is blurry or noisy, this estimate may be subject to error. The process of RST consists of using a reference structure to code the source radiation such that measurements can be used to compute estimations of source descriptors. A linear transformation such as the matrix r described above is “well-conditioned” if it has an inverse matrix T{circumflex over ( )}−1. This is the case if T is nonsingular and if the ratio of the largest eigenvalue or singular value of T to the smallest is not large. The matrix in the above example is well conditioned. [0091]
  • Multiplicative Diffuse RST [0092]
  • Multiplicative diffuse RST takes advantage of wave phenomena such as diffraction, refraction, scattering, diffusion, phase, frequency, etc. Visible projection RST neglects diffraction. Accordingly, under visible projections, a measurement point sums all source points based on their visibility from the measurement point. The visibility is proportional to the density of the ray through the reference structure from the measurement point to the source point. Under diffuse RST, radiation from source points may diffract around obstructions. In this case, the state of a measurement may depend on source points that are not visible from a source point. [0093]
  • As an example, visible projection corresponds to looking at an object through a forest. Light from the object may be sensed only if the object is unobscured by trees. However, if the same object were sensed through the forest by listening to sound made by it, the object can be sensed even if it is obscured by the trees because sound waves diffract around the trees. The signal diffracted by the trees still has some spatial sensitivity, i.e., the object may be detected through the trees differently depending on the object's and the detector's position relative to the trees). Using this diffuse spatial pattern to find or image the person through the trees is diffuse RST. [0094]
  • Diffuse RST does not function with rays, visible RST does. Neither diffuse not visible RST are likely to rely on Radon transforms or any other well known transformation. They can be designed to rely on some well-known mappings, which we have done in the case of the motion sensor reported below, but in general the mapping and its inverse may be an arbitrary linear transformation, such as transformations T shown above. A reference structure relying on such wave characteristics will generate more complicated mappings between the source and the measurement, and can thus yield more information about the source. Such mappings may be represented as [0095] m i = detector area source T i ( r , ρ ) S ( r ) r ρ ( 4 )
    Figure US20040095626A1-20040520-M00004
  • In FIG. 16, a [0096] reference structure 122 is disposed between a source 124 and a detector 120. A receiver pattern 126 is shown superimposed over the source 124. It should be noted that the receiver pattern 126 will not have rectilinear boundaries in the diffuse case.
  • Referring to FIG. 16, which is a schematic diagram of a RST system in which receiver patterns for diffuse propagation, analogous to those shown above for the 2×2 source of FIGS. [0097] 10-15, are shown. {right arrow over (r)} is a position variable over the source space 124. {right arrow over (ρ)} is used as in Equation 1 to refer to position on the sensor. The sensor is a finite area device that integrates the field over the area to produce a discrete number. {right arrow over (ρ)} in Eq. (4) integrates over the area of the detector.
  • T[0098] i({right arrow over (r)},{right arrow over (ρ)}) is the receiver pattern 126 for position {right arrow over (ρ)} on the ith detector. Ti({right arrow over (r)},{right arrow over (ρ)}) is a probability that the source signal radiated from point {right arrow over (r)} is detected by the ith detector. The order of integration in Eq. 2 is over the source variable {right arrow over (r)} on the inner integral followed by the detector position {right arrow over (ρ)} on the outer integral. The order of integration can be flipped, however and one achieves a simpler and mote intuitive equation by performing the integral over {right arrow over (ρ)} first, which yields
  • m i ={tilde over (T)} i({right arrow over (r)})S({right arrow over (r)})d{right arrow over (r)}  (5)
  • where {tilde over (T)}[0099] i({right arrow over (r)})=∫Ti({right arrow over (r)},{right arrow over (ρ)})d{right arrow over (ρ)} is the receiver pattern averaged over the detector area. One may interpret receiver pattern 126 to represent this average receiver pattern on the ith detector without loss of generality.
  • A mapping is a transformation from a distribution on one space to a distribution on another space. Eq. (4) is a mapping from the source distribution defined over {right arrow over (ρ)} in the modified Eq. (4) to the sensor state defined over the discrete positions indexed by i. Eq. (4) could describe either visible or diffuse RST, the difference between the two cases is that visible RST assumes that the receiver pattern is based on the “ray visibility” through the reference structure, diffuse RST allows the field radiated by the reference structure to diffuse through the structure. For example, the source field may consist of sound waves. The detectors can hear the sound even though the source may be visibly obscured because sound does not propagate in straight lines. However, the reference structure can still modulate the propagation of the sound waves such that a receiver pattern is produced and the source can be estimated. [0100]
  • T({right arrow over (θ)},{right arrow over (ρ)}) in Eq. 1 is not the same function as T[0101] i({right arrow over (r)},{right arrow over (ρ)}) in Eq. (4). In visible reference structure tomography, the mapping from the source to the sensor is based simply on the visibility through the reference structure from the sensor point to the source point. T({right arrow over (θ)},{right arrow over (ρ)}) is the visibility from the sensor point {right arrow over (ρ)} through the reference structure in direction {right arrow over (θ)}. Under diffuse RST, calculation of the mapping from the source to the sensors requires detailed analysis of field propagation through the reference structure. Ti({right arrow over (r)},{right arrow over (ρ)}) describes the results of this analysis between the source point {right arrow over (r)} and the sensor point {right arrow over (ρ)}, but we do not consider the details of this diffractive and refractive analysis here.
  • Eq. (1) is a triple integral because we use a simple algorithm for relating the reference structure to the mapping between the source and the sensor in visible RST (the visibility through the reference structure determines the mapping). Eq. (4) is only a double integral because we have not related the receiver pattern to the physical characteristics of the reference structure. Position in the source space in Eq. (1) is {right arrow over (r)}={right arrow over (ρ)}+αi[0102] {right arrow over (θ)}, the integral over α and θ in equation 1 corresponds to the integral over r in Eq. (4).
  • Referring to FIG. 16, the [0103] receiver pattern 126 shown on the source 124 is a probability density where signals or rays emitted from the corresponding source points are sensed by the detector 120 after modulation by the reference structure 122. As with the visible projection case, it is preferred to implement mappings that enable well-conditioned estimation of source variables. Note that the receiver pattern 126 is nonconvex (e.g., along lines through the receiver pattern one encounters multiple disjoint regions of high and low probability of detection. A convex region is a region such that all lines between points contained in the region are also contained in the region. A connected region is a region such that there exists at least one path between any two points in the region which is also in the region. A circular receiver pattern is both convex and connected. The cone associated with the receiver pattern for a pixel in a focal system is nonconvex but is connected.) Nonconnected receiver patterns may be desirable because they enable high-frequency and high resolution sensing without localizing the receiver pattern to a point.
  • While conventional optics, such as lenses, can be used as reference structures, these devices will not produce receiver patterns at the level of complexity obtained with multiple dimensional modulations. Reference structures which are modulated non-monotonically in multiple dimensions are necessary to produce complex nonconvex receiver patterns. Non-monotonic modulation is a modulating signal that contains multiple local minima and maxima (it does not increase or decrease monotonically). In the context of simple thin lens design, a lens does not modulate a wavefield in depth and modulates the optical path length monotonically up to some maximum and then monotonically back down to a minimum. A reference structure will include multiple minima and maxima in the modulating refractive, diffractive or absorptive structure along lines through the reference structure. [0104]
  • Additive RST [0105]
  • Tomographic imaging systems can be based on particular models for the source/detector mapping. Under the multiplicative RST model discussed above, it is assumed that projections consist of a linear modulation on propagation through the source and the reference structure individually, but that the reference structure modulates the source signal multiplicatively. This model is appropriate when the reference structure strongly modulates the source signal, as when the reference structure consists of obscuring structures or apertures in an opaque mask. This is the case in the 2×2 example of FIGS. [0106] 11-15.
  • Visible multiplicative and additive RST are based on different approximations to the transmission model [0107] t ( ρ , θ ) = exp ( - ray extent [ T ( ρ + α i θ ) + S ( ρ + α i θ ) ] α ) , ( 6 )
    Figure US20040095626A1-20040520-M00005
  • where t({right arrow over (ρ)},{right arrow over (θ)}) is the signal transmitted by the combination of the source and reference structure to the measurement point {right arrow over (ρ)} along the ray direction {right arrow over (θ)}. Across the detector, the signal resulting from t({right arrow over (ρ)},{right arrow over (θ)}) is [0108] m i = detector area detector acceptance angle t ( ρ , θ ) θ ρ ( 7 )
    Figure US20040095626A1-20040520-M00006
  • The multiplicative approach uses the separation [0109] t ( ρ , θ ) = exp ( - ray extent [ T ( ρ + α i θ ) ] α ) exp ( - ray extent [ S ( ρ + α i θ ) ] α ) ( 8 )
    Figure US20040095626A1-20040520-M00007
  • and defines [0110] T ~ ( ρ , θ ) = exp ( - ray extent [ T ( ρ + α i θ ) ] α ) ( 9 )
    Figure US20040095626A1-20040520-M00008
  • as the visibility of the reference structure. The approximation [0111] exp ( - ray extent [ S ( ρ + α i θ ) ] α ) 1 - ray extent [ S ( ρ + α i θ ) ] α ( 10 )
    Figure US20040095626A1-20040520-M00009
  • the [0112] yields equation 1. The additive approach makes the alternative approximation t ( ρ , θ ) 1 - ray extent [ T ( ρ + α i θ ) + S ( ρ + α i θ ) ] α ( 11 )
    Figure US20040095626A1-20040520-M00010
  • The additive approach as represented by Eq. (11) is not useful unless the reference and source signals are nonlinearly combined as in Eq. (13) below. This occurs, for example, when a coherent field is detected by a square law detector. [0113]
  • The most likely example of a weak modulation reference structure is a volume hologram based on 3D absorption or phase modulation. 3D imaging using volume holograms is condensed in, for example: G. Barbastathis, M. Balberg, and D. J. Brady, “Confocal microscopy with a volume holographic filter,” [0114] Optics Letters, vol. 24, pp. 811-813, 1999; G. Barbastathis and D. J. Brady, “Multi-dimensional tomographic imaging using volume holography,” Proceeding of the Ieee, vol. 87, pp. 2098-2120, 1999; or G. Barbastathis and D. J. Brady, “Volume holographic imaging of three-dimensional objects,” Proceedings of SPIE—The International Society for Optical Engineering, vol. 3633, pp. 170-181, 1999, all of which are incorporated herein by reference in their entirety. The difference between RST and previous work is that the hologram is designed to measure tomographic projections based on dbResponse.9 rather than focal images.
  • In some cases, an additive reference structure, where the reference structure weakly modulates the source signal, may be preferred. Weak modulation is attractive when the additive mapping, represented by Eq. (12) below is desired. This mapping is attractive when the detector signal is proportional to the intensity of the radiation field, as represented in Eq. (13) below. [0115]
  • In the case of an additive reference structure, the reference structure signal will add to the source, yielding the mapping: [0116] m i detector area detector acceptance angle ray extent [ T ( ρ + α i θ ) + S ( ρ + α i θ ) ] α θ ρ ( 12 )
    Figure US20040095626A1-20040520-M00011
  • i[0117] {right arrow over (θ)} represents a unit vector pointing in the direction of the angle vector {right arrow over (θ)}. As illustrated in FIG. 17, {right arrow over (θ)} is a simple angle in the plane in 2D, but becomes a two dimensional position on the unit sphere in 3D (in mathematical terms {right arrow over (θ)} is a point in Sn, the n dimensional sphere). It is not really necessary to make a distinction between i{right arrow over (θ)} and {right arrow over (θ)}, but in common usage in 3D {right arrow over (θ)} might be the ordered pair (θ,φ) and i{right arrow over (θ)}=sin θ cos φix+sin θ sin φiy+cos θiz.
  • The geometry of additive and multiplicative modulation is the same and the difference lies in the reference structures modulation properties and the detection model illustrated in Eq. (13) below. Multiplicative RST is probably much more useful for optical applications, but additive RST is more useful in extending the utility of x-ray crystallography. [0118]
  • The additive mapping is most useful in cases where the source is considered as a coherent field and the detection process is nonlinear, as is the case with intensity detectors. In this case the measurement can be described as [0119] m ~ i = m i 2 = μ R + μ S 2 = μ R 2 + μ S 2 + μ s * μ R + μ R * μ S ( 13 )
    Figure US20040095626A1-20040520-M00012
  • where μ[0120] R is the field due to the reference structure and μS is the field due to the source. If the reference structure is designed to produce a scattered or radiated field of a known frequency or known basis, it is possible to filter on this frequency or basis to isolate μS from the measurement, essentially achieving self-referencing structural holography.
  • Some of the advantages of RST are that it enables reconstruction of sources without scanning and that it may allow efficient reconstruction of sources with fewer measurements. Current tomographic and holographic systems use a priori knowledge of the probing field, as in illumination or measurement from a particular direction or sequence of directions in tomography or interference with a reference wave in holography. RST uses the reference structure to eliminate the need for special illumination and for scanning sequences. In this sense, reference structure modulated measurement of a source may be termed self-referencing. In particular, the coherent (additive RST case) or incoherent (multiplicative RST case) field radiated by the source provides its own interferometric (additive case) and/or projective (multiplicative case) reference fields by structuring source radiation patterns. Because a source in a reference structure modulated measurement does not need to be scanned as a function of time, the source space might also be termed “self-revealing.”[0121]
  • As an example, many diffractive scattering systems rely on the Fourier transform relationship between far-field scattering patterns and a near-field structure. In this case, it may be assumed that μ[0122] S={tilde over (S)}({right arrow over (k)}) and μR={tilde over (R)}({right arrow over (k)})ej{right arrow over (k)}·{right arrow over (d)} where {right arrow over (d)} is the separation between the center of the reference structure and the center of the source and {right arrow over (k)} is a wave number proportional to the direction between the source and the detector and the illumination wavelength. It is then possible to filter on a frequency proportional to {right arrow over (d)}, if {right arrow over (d)} is less than the coherence length of the illumination. For example, this relationship forms the basis for heavy atom decoding in x-ray crystallography. On molecular scales, the coherence length would have to extend to several nanometers for molecular reference structure tomography. For example, see: E. delaFortelle and G. Bricogne, “Maximum-likelihood heavy-atom parameter refinement for multiple isomorphous replacement and multiwavelength anomalous diffraction methods,” in Macromolecular Crystallography, Pt A, vol. 276, Method of in Enzymology, 1997, pp. 472-494; R. W. Grosse-Kunstleve and A. T. Brunger, “A highly automated heavy-atom search procedure for macromolecular structures,” Acta Crystallographica Section D-Biological Crytallography, vol. 55, pp. 1568-1577, 1999; U. Heinemann, G. Illing, and H. Oschkinat, “High-throughput three-dimensional protein structure determination,” Current Opinion in Biotechnology, vol. 12, pp. 348-354, 2001; or T. V. Ohalloran, S. J. Lippard, T. J. Richmond, and A. Klug, “Multiple Heavy-Atom Reagents for Macromolecular X-Ray Structure Determination—Application to the Nucleosome Core Particle,” Journal of Molecular Biology, vol. 194, pp. 705-712, 1987, all of which are incorporated herein by reference in their entirety.
  • The measurement term of particular interest in this case is [0123]
  • μR *μS ={tilde over (R)} *({right arrow over (k)}){tilde over (S)}({right arrow over (k)})e −j{right arrow over (k)}·{right arrow over (d)}  (14)
  • In the ideal case, where {tilde over (R)}({right arrow over (k)}) is equal to a constant, the measurement is simply proportional to {tilde over (S)}({right arrow over (k)})e[0124] −j{right arrow over (k)}·{right arrow over (d)} and the source distribution can be obtained by filtering about the carrier {right arrow over (d)}. (Filtering in this context refers to considering only a window centered on {right arrow over (d)} as representative of the source density S({right arrow over (r)}) in the inverse Fourier transform of the measurement.) To obtain this situation, the reference structure must be a very strong point scatterer. Since any point scatterer is unlikely to scatter sufficient power to enable reconstruction, heavy atom methods use diffuse arrays of point scatterers to generate a reference signal. In heavy atom methods a discrete set of strong point scatterers is embedded in a target molecule. In this case, R ~ ( k ) = strong scatterers j k · d i .
    Figure US20040095626A1-20040520-M00013
  • Formally, one estimates {tilde over (S)}({right arrow over (k)})e[0125] −j{right arrow over (k)}·{right arrow over (d)} in this case by dividing Eq. (14) by {tilde over (R)}({right arrow over (k)}), although other methods are likely in practice. Additive reference structure tomography is a generalization of heavy atom methods to weaker and more distributed scatterers and to novel sensor modalities in optics, acoustics and other propagating field modalities. In the case of x-ray crystallography, additive RST might consist, for example of binding unknown target molecules to known reference structure molecules to enhance x-ray determination of molecular structure.
  • An example of a benefit of RST is for scan-free multi-dimensional imaging, as shown in the simple 2×2 sensor example of FIGS. [0126] 11-15, where RST can achieve multi-dimensional source reconstruction without the spatial and temporal scanning associated with conventional tomography. In contrast to conventional tomography, RST can sort and filter radiation to obtain all projections in a single time step. The reference structure determines the spatial sampling rate and limits the resolution of the source reconstruction, but this limit is not necessarily worse than the restrictions imposed by conventional sampling.
  • As noted above, RST is used to estimate sources on multi-dimensional spaces without scanning, to estimate sources on unusual bases, such as modal densities as a function of time and to estimate sources using fewer measurements than naïve system design might suggest. To understand the sensor efficiency application, consider a source distributed over N[0127] p discrete spatial sampling points or resolution cells. The dynamic range, meaning the number of different discrete values that the source distribution in each cell can assume, is Dp. If the state of each cell is independent, the number of different states this source can assume is Dp N p . Now suppose that we wish to estimate this source using Nm measurements of dynamic range Dm. The number of measurement states exceeding the number of possible source states is a necessary but not sufficient condition for discrimination of each source state. For the source described above this requirement implies that NmlogDm≧NplogDp. Conventional tomographic and imaging systems satisfy this requirement by making at least on measurement per resolution cell or pixel.
  • Reference structures enable the [0128] receiver pattern 126 illustrated in FIG. 16 to be implemented with multiple spatial resolutions and to combine source points in coded patterns. Receiver pattern programming enables RST to improve sensor efficiency over conventional systems in cases where the state of different resolution cells are not independent. Few sources exhibit such independence in practice, in general, the state of a given cell is influenced by the states of adjacent cells. For example, in an image of a face, the presence of an eye in one region of the image implies that another eye will appear elsewhere in the image. The lack of independence of different regions of the image is the basis of image compression algorithms, which commonly reduce the raw image data by 1-3 orders of magnitude. Reference structure tomography can be used to implement such compression directly in the physical sensor, so that the amount of data sensed and processed is substantially reduced with little or no cost to system performance.
  • It is possible to quantify the efficiency of a sensor system by defining the number of possible sources N[0129] S. For a random source, NS=Dp N p , but for common sources NS is dramatically less than this limit. We define the efficiency of a given sensor system to be ξ = log N s N m log D m ( 15 )
    Figure US20040095626A1-20040520-M00014
  • and say that a sensor is efficient if ξ1. In general the number of sources may be limited in a variety of ways. For example, the source may consist of a discrete set of objects or patterns. To illustrate the use of RST to achieve efficient sensing, we assume here that the source consists of one point radiator occupying one of N[0130] p resolution cells. This situation often occurs in motion imaging systems, in which at most one resolution cell is changing as a function of time. The problem is easily generalized to any finite number of occupied cells, however. A single occupied cell produces at most Np states. A conventional sensor with Nm=Np yields the sensor efficiency ξ = log N p N p log D m 1.
    Figure US20040095626A1-20040520-M00015
  • Under certain configurations, it is possible to use reference structures to shape the spatial response pattern of, for example, a microphone for multi-dimensional acoustic imaging or the response of a pyroelectric detector for infrared tracking and identification. Such a sensor may be capable of finding a single source among N[0131] p possible locations in log2(Np) measurements, thus achieving ξ=1. An RST measurement on such a source measurement consists of a weighted sum of the resolution cell densities of the form mi=Σ j=1N p αijsj, where s is the density in the jth cell and αijε{0,1}. {sj} is the discretized source density and {αij} is the discretized receiver pattern created by the reference structure.
  • Given that only zero or one resolution cells are active, the following sequence of measurements determines the density of the occupied cell. α[0132] ij=1 measures the total density of the source, which in this case is the density of the occupied cell. For i=2 to log2Np+1: α ij = { 1 odd ( floor ( j 2 i - 2 ) ) 0 even ( floor ( j 2 i - 2 ) ) ( 16 )
    Figure US20040095626A1-20040520-M00016
  • The address of the occupied cell is the binary number [0133] ( m log 2 N p m 1 ) ( m log 2 N p - 1 m 1 ) ( m log 2 N p - 2 m 1 ) ( m 3 m 1 ) ( m 2 m 1 ) ( 17 )
    Figure US20040095626A1-20040520-M00017
  • Thus log[0134] 2Np+1 measurements are required if Np is a power of two, and log2Np+2 measurements are required if Np is not a power of two.
  • Multi-Dimensional Source Reconstruction Using RST
  • When multi-dimensional imaging, rather than sensor efficiency, is the goal, RST systems are designed such that the number of measurement points is equal to or greater than the number of source resolution cells. FIG. 17 illustrates a geometry for analysis of spatial scaling vs. resolution for two-dimensional RST. In the analysis, the source [0135] 128 is described by a density function s(x,y). As stated above, computed tomography imaging is commonly implemented by digital inversion of a sampled set of ray projections of the density function through the object to be imaged. In conventional implementations, these projections are sampled using a temporal sequence of illumination and reconstruction geometries. These projections are measured and interpreted in the context of a source and illumination model.
  • The most common interpretation model assumes that each ray projection is a simple integral of the source density along the ray path. Such a model can suggest the assumption that s(x,y) represents a florescence signal from a point located at (x,y), and that the florescence from each point source radiates out from the source at an angle θ without interaction with the rest of the source. Under this assumption, projections through the source are of the form [0136]
  • g(l,θ)=∫s(l cos θ−α sin θ,l sin θ+α cos θ)  (18)
  • As illustrated in FIG. 17, l is the shortest distance from the origin to the ray and θ is the angle the ray forms with respect to the x axis. [0137]
  • Taking the Fourier transform of Eq. (18) with respect to l results in [0138] G ( u l , θ ) = S ( u , v ) j 2 π ( u ( l cos θ - α sin θ ) + ( l sin θ + α sin θ ) v ) - j 2 π v i l u v l α = S ( u = u l cos θ , v = u l sin θ ) ( 19 )
    Figure US20040095626A1-20040520-M00018
  • where S is the Fourier transform (FT) of s, meaning that the FT of a projection is the slice of the FT of the source. If “l” space is sampled uniformly along an aperture of length R[0139] s with sampling period Δl, then the resolution along slice in the reconstructued Fourier space G is Δul=2π/Rs and the range in the (u,v) Fourier space spanned by G is U=4π/Δl. Measurements of g(l,θ) are also sampled in angle. Many sampling algorithms may be considered, but we can estimate the resolution and span obtained in a given geometry by assuming that g(l,θ) is sampled uniformly in θ with a sampling period Δθ. In this case, samples in the Fourier space are polar sampled on a disk of radius 2π/Δl. The maximum sampling period on the disk is the greater of 2π/Rs and 2πΔθ/Δl. To obtain quasi-uniform 2D sampling, we set Δθ=Δl/Rs.
  • Many algorithms have been developed for mathematical inversion from the projections g(l,θ) to estimate the source distribution s(x,y). Formally, one imagines inversion based on using the projection slice theorem (Eq. 19) for various θ to fill the estimated Fourier space followed by inverse Fourier transform. The reconstruction space will produce as many sample points as there are projections. In conventional computed tomography systems these projections are obtained by illuminating the source with parallel bundles of rays incident at sampled angles θ. Alternative bundles, such as fan-beam and cone-beam variations may be used, but all require sequential scanning. [0140]
  • A reference structure can eliminates to scan in obtaining tomographic projections. Referring to FIG. 18, a radiation source lies within a [0141] circle 136 of radius Rs, which lies within a measurement circle 130 of radius Rm. Rays 132 passing through the source space and incident on the measurement circle are also illustrated. The rays are shown in bundles for values of l spaced by Δl relative to each other, where l is defined as in FIG. 17. The angular displacement from one bundle of rays to the next is Δθ, where θ is also defined as in FIG. 17. The separation Δl between the sampling rays corresponds to the pixel spacing on the measurement circle. Note that the sampling rays cross in the region adjacent to the source. We can insure that the only one ray is incident on each sensor point by setting tan ( R s R m ) Δ θ .
    Figure US20040095626A1-20040520-M00019
  • Assuming that R[0142] m□ Rs, we find that the rays can be separated so long as Rm>Rs 2/Δl. As discussed above, Δl is the approximate resolution of the tomographically reconstructed estimate of the source. Assuming that signals propagate from source fluorescence or scattering points in straight rays, a reference structure in the form shown in FIG. 19 will sample tomographic projections on the appropriate Δl, Δθ lattice for reconstruction using Eq. (19). The reference structure would consist of the circular center well 137 of radius Rs surrounded by a circular absorbing medium 131 of radius Rm. Holes would be cut through the absorbing medium on spacings of Δl. Bundles of holes spanning the Rs radius of the inner circle would be formed as shown in the figure. These bundles would be spaced on angular displacements of Δθ. Detectors at the ends of each hole would measure the transmittance or radiance of the source along the corresponding ray path, thus sampling g(l,θ) as in Eq. (18).
  • FIG. 19 illustrates that it is possible to build a 2D reference structure that can reconstruct a 2D source tomographically without scanning as long as the radius of the reference structure is greater the square of the radius of the source divided by the spatial resolution of the reconstruction. This argument can be extended to higher dimensions by extending ray projections as in FIGS. 18 and 19. For a three dimensional source, for example, one may consider layered slices of the 2D ray bundles as shown in FIGS. 18 and 19, fan-beam or cone-beam sampling geometries or other ray clusters. FIG. 20 shows a reference structure analogous to FIG. 19 for a 3D source. The structure consists of layers of [0143] 2D reference structures 138 spaced by Δz. As in FIG. 19, the black rays in the figure would be holes between detectors on the measurement cylinder. The outer measurement cylinder 132 is of radius Rm. The inner cylinder 139 would be a hollow container for the source under measurement of radius Rs Separation of the rays requires that the hole diameters be less than Δz and Δl and that Rm>Rs 2/Δl, as above.
  • An alternative approach to constructing a 3D reference structure which yields both better light efficiency and a smaller reference structure relies on the 3D radon transform. Under this approach measurement points on the surface of the reference structure measure the signal radiated by planes of the source. The measured function takes the form [0144]
  • g(l,{right arrow over (θ)})=∫s(li {right arrow over (θ)} −αi {right arrow over (θ)}⊥1 −bi {right arrow over (θ)}⊥2)dαdβ  (20)
  • As used previously in this description i[0145] {right arrow over (θ)} s a unit vector in the direction of the angles {right arrow over (θ)}(θ,φ) on the unit sphere. i{right arrow over (θ)}⊥1 and i{right arrow over (θ)}⊥2 are unit vectors in three dimensions orthogonal to i{right arrow over (θ)} and to each other. The geometry of the integration of Eq. (20) is illustrated in FIG. 21. A detector at the measurement point 133 with an unobstructed view of the plane spanned by i{right arrow over (θ)}⊥1 and i{right arrow over (θ)}⊥2 a distance l away from the origin measures g(l,{right arrow over (θ)}) as in Eq. (20). Again taking the Fourier transform with respect to l, we find the 3D version of the projection slice theorem G ( u l , θ ) = S ( u ) j 2 π l ( u · i θ - u i ) j 2 π α u · i θ 1 j2πβ u · i θ 2 u v l α β = S ( u = u l i θ ) ( 21 )
    Figure US20040095626A1-20040520-M00020
  • One uses a reference structure to measure g(l,{right arrow over (θ)}) by mapping leaving conic sections of the form shown in FIG. 21 unobscured for measurement points on the surface of a sphere of radius R[0146] m. The cross section of the point of the cone at the measurement point determines the reconstruction resolution Δx. If one uses a sampling scheme that full populates the outer sphere with sampling points, equality between the number of data points on the measurement sphere and the number of reconstructed source points will require Rm 2>Rs 3/3Δx. Many sampling schemes may approach this limit.
  • As an example of reference structure design for 3D Radon transform sampling, we assume that we sample points on the measurement sphere at the following angular coordinates: [0147] φ i j m = j Δ x R m sin ( θ i m ) j = 0 , 1 , 2 2 π R m sin ( θ i m ) Δ x θ i m = i Δ x R m i = 0 , 1 , 2 π R m Δ x ( 22 )
    Figure US20040095626A1-20040520-M00021
  • φ[0148] ij and mθ are angular positions on the measurement sphere. This strategy produces approximately 4πRm 2/Δx samples. We assign samples on the measurement sphere to samples in the 3-dimesional (l,{right arrow over (θ)}) space on which we sample g(l,{right arrow over (θ)}) according to some algorithm. For example, setting φ i j = φ i j m θ i = θ i m + cos - 1 ( R s R m φ i j m 2 π ) 0 < θ i < π l = R s φ i j m 2 π ( 23 )
    Figure US20040095626A1-20040520-M00022
  • would sample using a reference structure formed around clear polygons as shown in FIG. 22. The transmissive region formed in the figure would be a non-absorbing channel through which source radiation over a plane of the source is visible to a detector at an outer vertex. FIG. 23[0149] a shows channels for Δx=Rs/100, which produces over a million sample points on a measurement sphere approximately 6 radii out from the source volume.
  • The sampling scheme described by Eq. (23) produces a radially sampled Fourier space of {right arrow over (u)} over S({right arrow over (u)}). Samples are obtained at evenly spaced radii along rays at the sampled angular values. The Fourier space sample frequency is 1/R[0150] s and the radius of the band volume covered is approximately 1/Δx.
  • Photon Efficiency and RST
  • While Radon transform-based reference structures illustrate that reference structures can be constructed to enable scan-free multi-dimensional sensing, the light efficiency of these structures is relatively poor. In the ray based structures the power detected by the sensors is only [0151] 1 N p
    Figure US20040095626A1-20040520-M00023
  • of total radiated energy, with the remainder of the energy absorbed by the reference structure. As used previously, N[0152] p is the number of resolution elements in the source. The 3D planar sampling reference structures obtain somewhat better performance, detecting approximately N p - 1 3
    Figure US20040095626A1-20040520-M00024
  • of the radiated power. [0153]
  • Reference structures can be designed to provide multiplex coverage with a larger system response function, which improves photon efficiency. Radon based transforms described designed such that only {square root}{square root over (N[0154] p)} source cells are detected by each detector in the 2D case and N p 2 3
    Figure US20040095626A1-20040520-M00025
  • are detected in the 3D case. In general, signals from all the source cells are radiated to each detector element and unwanted signals must be obscured or absorbed to enable Radon-based coding. One could overcome the poor photon efficiency implicit in this absorption by increasing the number of detector elements, perhaps by making the reference structure itself from detector elements, or by designing the reference structure to be as transmissive as possible. [0155]
  • Alternative transformations, such as Hadamard or Fourier-transform based schemes might be implemented on the source space to improve photon efficiency. Hadamard and Fourier systems measure approximately half of the power incident on each detector element. We describe these approaches further below when considering detailed implementations. [0156]
  • IMPLEMENTATION EXAMPLES Transmission Mask Based Reference Structures
  • For such stacked reference structures, a point source, as viewed through the masks, can be analyzed as follows. The transmittance of the i[0157] th mask is ti(x), as shown in the schematic representation of FIG. 23b, where the source 162 projects a signal ray 168 through a series of masks 164 to a detector 166. For the ith mask, the signal 168 received on the detector 166 as a function of position along the detector 166 surface is m ( x ) = i t i ( x - l i ( x s - x ) ( R - l i ) ) ( 24 )
    Figure US20040095626A1-20040520-M00026
  • where R is the range to the source. As an example, let t[0158] i(x)=1+cos(Kx) and li=iΔ. Then m ( x ) = i ( 1 + cos ( K x ( 1 + i Δ R ) - K i Δ x s R ) ) ( 25 )
    Figure US20040095626A1-20040520-M00027
  • FIG. 24 is a graph that illustrates the transmission of the i[0159] th mask for Kxs=100, R=1000 delta with 10 masks m(x) for Kx=1:100. FIG. 25 is a transmission graph of the same system with the number of masks increased from 10 to 100. FIG. 26 is a transmission graph utilizing the same parameters used to generate the graph of FIG. 25, but where the cosine function has been replaced with a rectilinear function. FIG. 27 is a transmission graph of an RST system with x fixed at 0 and a function of xS for a fixed x. FIG. 28 is a transmission graph of an system with x fixed at 0, and with KxS =1:100, R=1000 delta and 10 masks m(x) for Kx=0. FIG. 29 is a transmission graph of an RST system, with Kxs=−100:100, R=100 delta and 10 masks m(x) for Kx=0.
  • Plane Wave Model for a Layered Mask Reference Structure [0160]
  • FIG. 30 is a schematic diagram of a plane wave motion sensor model, including a motion sensor as a mapping from plane waves onto measurements. In FIG. 30, a plane [0161] wave incidence signal 200 is received by a reference structure 204 at an angle of θ to the normal 202 to the surface of the reference structure 204. Sensor points 206 are disposed behind the reference structure 204 and configured to receive any signal from the reference structure 204. In this configuration, it is preferred that the reference structure 204 consists of N layers of transmittance masks and that the incident signal 200 is f(θ). The incidence signal 200 detected on the ith sensor element is m i = n = 1 N pixel_i f ( θ ) t n ( x - n l sin θ ) θ x ( 26 )
    Figure US20040095626A1-20040520-M00028
  • where 0≦t[0162] n(x)≦1.
  • For the example, where [0163] t n ( x ) = m α m rect ( x - m Δ Δ )
    Figure US20040095626A1-20040520-M00029
  • where α[0164] m
    Figure US20040095626A1-20040520-P00900
    {0,1}, then m i = n = 1 N pixel_i m f ( θ ) α n m rect ( x - m Δ - n l sin θ Δ ) θ x ( 27 )
    Figure US20040095626A1-20040520-M00030
  • For picking the transmittance functions, it is assumed that it is possible to model the [0165] reference structure 204 based on a reduced Hadamard S matrix (eliminating the dc row and column). Then, 31 angles build the connection matrix shown in FIG. 31.
  • As an example of an embodiment of a reference structure for plane wave detection designed to detect a source consisting of 8 angularly resolved picture elements on a side (64 total), consider a sensor plane which consists of an 8 by 8 grid of pixels. Since any measurement pixel could measure any incident ray, the first screen of the layered reference structure must be of [0166] size 16 by 16. FIGS. 32 and 33 are graphs that show the 3D output of such a configuration by sampling longitudinally with masks.
  • FIG. 34 is a schematic diagram of mask and sensor geometry and shows a [0167] transparency 232 and a sensor assembly 236, where the sensor assembly 236 includes and 8×8 array of sensors 238 of D×D dimensions. Ray projections 234 are received by sensors 238 from the transparency 232. As illustrated in FIG. 34, the distance from the sensor 236 plane to the entrance aperture of the transparency 232 must be sufficient to separate rays 234 onto adjacent channels or sensors 238, i.e. L>Δ/Δθ.
  • In this example, the dimensions D of each [0168] sensor 238 correspond to pixel spacing of 1 mm. If 1 cm resolution at a range d of 1 meter is desired, then Δθ=0.01. If an 8×8 grid with the center pixel removed is preferred, the reference structure with transparency 232 will correspond to a 16 mm square input aperture and a reference structure thickness of 10 cm. FIG. 35 is a grid showing how the measurement pixels are numbered in the sensor assembly 236 of FIG. 34. The grid position as a function of pixel number for an n×n array can be determined, for example, with the computer program listed in Appendix A.
  • Connectivity can be assigned using the algorithm listed in Appendix A. Different views of the connectivity pattern of the reference structure for the reduced Hadamard connectivity are illustrated in FIGS. 36[0169] a-36 c at various angles of view where each connectivity pathway 242, runs between a reference structure entrance 240 and a reference structure exit 244.
  • FIG. 37 is a graph which shows the singular value spectrum associated with this interconnection pattern. [0170]
  • FIG. 38 is a schematic diagram of [0171] reference structure pipes 246 with a length of L, a width of d and an aperture angle θ. A ray 250 may travel through the pipe 246 off-axis up to an angle θ. For the example in FIG. 38 there are 100 micron pixels configured to detect over a θ=10 degree range, meaning that each ray covers 2.5 degrees.
  • FIG. 39 is the reduced S matrix. FIG. 40 is the connectivity pattern for the configuration shown in FIG. 38. In FIG. 40, the [0172] connectivity paths 254 lie between the entrance 252 to the reference structure and the exit 256 to the reference structure. FIGS. 41a-41 k are cross-sectional schematic views of a set of masks, in order from an input side to an output side where each mask 258 a-258 k has multiple apertures 260. Each mask 258 a-258 k may be stacked or layered to form the corresponding reference structure. As shown in FIGS. 41b-41 i, the apertures 260 of masks 258 b-258 i may overlap other apertures 260.
  • Mask Thickness vs. Resolution [0173]
  • For a reference structure with a thickness L for each mask, and an opening of size δ for each aperture in the mask, the angular resolution is Δθ=2δ/L. The number of angular resolution elements along each transverse dimension is N≦A/δ, where A is the full transverse aperture. The transverse angular range observed is Ω. Using N≦Ω/Δθ, L≧2 A/Ω is obtained for the thickness of the reference structure for a given aperture size and angular range. [0174]
  • FIG. 42 is a schematic of a system for imaging a 4×1 LED array using a reference structure. In FIG. 42, the [0175] reference structure 263 consisting of a solid opaque block with transparent paths or pipes 264 therethrough. The transparent pipes 264 run from the input aperture 262 of the reference structure 263 to the exit aperture 266. Such a reference structure 263 can be made by fabrication techniques including selective laser sintering, stereolitograph and holography. In the example of FIG. 42, the reference structure 263 is designed to accept plane waves at the reference structure input aperture 262 at 4 angles of approximately −10°, −5°, 5°, 10°. The reference structure 263 consists of 10 of the transparent pipes 264 of 1 mm diameter, each oriented at the aforementioned angles at the input 267 of the reference structure 263. Each of these pipes 264 goes to one of the four sensors in a sensor array space close to the reference structure 263. Each sensor may receive a combination of these pipes 264 as shown in FIG. 42.
  • Referring to FIG. 43, a schematic of an experiment is shown, where a [0176] sensor array 272 receives light from LEDs 267 of an LED array 268 through a reference structure 270. The LED array 268 was used as the source to characterize the matrix representing the reference structure 270. The LED array 268 consisted of 4 LEDs 267 spaced 5 cm apart, and was placed 50 cm from the reference structure 270. A 12 bit cooled CCD camera was used as a sensor array 272. Each sensor was chosen to be a group of 60×50 pixels (pixel size 4.65×4.65) on the cameras imaging surface.
  • FIGS. 44[0177] a-44 o show the patterns 274 a-274 o transmitted to the camera by the reference structure 270 for different combinations of lighted LEDs 267.
  • The pixel intensities which exceed a pre-selected threshold are added together to get a total intensity value for each sensor of the [0178] sensor array 272. The sensor intensity map for different combinations of lighted LEDs 267 in the LED array 268 is shown in Table 1 below.
    TABLE 1
    LED pattern Sensor intensity pattern (S1, S2, S3, S4)
    0001 0 881874 330926 0
    0010 1197105 0 1743120 0
    0011 1194541 887534 2080042 0
    0100 538590 733136 623103 257955
    0101 537086 1638702 967603 254241
    0110 1723678 730187 2347902 256039
    0111 1515366 1441309 2353960 217363
    1000 0 0 301831 487603
    1001 0 894262 663928 481717
    1010 1197316 0 2070979 484394
    1011 1201341 907061 2410049 486254
    1100 546877 735763 953365 777133
    1101 550028 1656286 1303013 779551
    1110 1611061 683155 2483884 719698
    1111 1405834 1332357 2432880 617232
  • The intensity values corresponding to the [0179] LED pattern 1000, 0100, 0010, 0001 constitute the columns of the matrix, which characterize the reference structure 270 represented by the matrix. H = [ 0 538590 1197105 0 0 733136 0 881874 301831 623103 1743120 330926 487603 257955 0 0 ]
    Figure US20040095626A1-20040520-M00031
  • The inverse of this matrix is used to invert the sensor intensities to get the source pattern comprising whether each [0180] LED 267 is “on” or “off.” The inverted values obtained are shown in Table 2 below.
    TABLE 2
    LED pattern Inverted values
    0001 0, 0, 0, 1.0
    0010 0, 0, 1.0, 0
    0011 0, 0, 1.0, 1.0
    0100 0, 1.0, 0, 0
    0101 0, 1.0, 0, 1
    0110 0, 1, 1, 0
    0111 0, 0. 9, 0.9, 0.9
    1000 1.0, 0, 0, 0
    1001 1.0, 0, 0, 1.0
    1010 1.0, 0, 1.0, 0
    1011 1.0, 0, 1.0, 1.1
    1100 1.0, 1.0, 0, 0
    1101 1.0, 1.0, 0, 1.0
    1110 1.0, 0.9, 0.9, 0
    1111 0.8, 0.8, 0.8, 0.8
  • The values are corrected to one decimal place. [0181]
  • 3D Reference Structures for Motion and Tracking Sensors [0182]
  • Infrared motion sensors are widely used in security and control applications and can be improved with the addition of a reference structure. Conventional motion sensors consist of two or more detectors operating in differential mode, where the measurement of each detector is referenced against the other. As a source moves through the receiver pattern of these detectors, changes in the states of the detectors relative to one another occur. The receiver pattern of such a detector is a 3D density map of the probability that a photon emitted at a point in space hits the detector. [0183]
  • Referring to FIG. 45, a schematic of an example of a receiver pattern for a conventional 2-element motion sensor is shown. In FIG. 45, two [0184] detectors 276, with optics 277 exhibit a receiver pattern with two detection areas 278 and 280. FIG. 46 is a schematic in which the two detectors 276 with optics 277 have a receiver pattern with three detection areas 282 interposed between 3 non-detection areas 284.
  • An embodiment of the invention expands the utility of motion sensors to include object recognition and tracking capabilities by making receiver patterns more complex. Receiver patterns of conventional motion detectors are shaped to discriminate, for example, between people and animals or vehicles. These patterns are convex, however, meaning that the receiver pattern is a beam, shaped mostly like the target. [0185]
  • For a motion sensor, the receiver patterns are structured such that the temporal signature on the detectors as the source moves through the field of view corresponds to the range to the source. Since the receiver patterns will vary with wavelength of the radiation emitted by or which otherwise propagates from the source, the temporal pattern may also give a spectral signature for the source, allowing for source identification as well as tracking. Complex receiver patterns are particularly attractive as embedded processors and wireless communications ports are added to motion sensors. [0186]
  • FIG. 47 is an example of a receiver pattern on a wireless array. In FIG. 47, two [0187] detector arrays 288 and 290 with corresponding optics 289 and 291 provide a receiver pattern forming a spatio-spectral grid 286 with intersecting detection areas 285 to detect the presence and motion of a source 292. The spatio-spectral grid 286, which the source 292 moves through on the array, will produce a temporal signal on the array of detectors 288 and 291 which allows for the determination of a characteristic of the source, and its position and velocity. An embodiment of the invention in which the motion sensor is enclosed within a reference structure may be used as a densely deployed and integrated component of a heterogeneous environment of sensors such as acoustic motion sensors, seismic sensors, video and IR cameras, etc.
  • Preferred embodiments of reference structures include 3D and 2D non-monotonic structures. As discussed above, such, example structures may have transmittance functions based on Hadamard transforms (i.e., Hadamard matrices). Furthermore, reference structures may be implemented using materials that modulate absorption, permittivity, impedance, florescence or other field propagation properties. [0188]
  • Another example of a beneficial use of RST is for phase sensitive crystallography. For phase sensitive crystallography, RST can be used to sense the phase of multi-dimensional scattered fields. This sensitivity may be useful in macromolecule structure analysis. It may be preferable to implement RST by binding a well-characterized reference molecule to an unknown source molecule in a crystal structure. The reference scattering enables the phase sensitive Fourier representation of the unknown molecule to be determined. [0189]
  • FIG. 48 is a perspective ghost view of a [0190] reference structure 154, in accordance with one embodiment of the present invention. The reference structure 154 has an input side 160 and an output side 150. The input side 160 includes multiple input windows 158. The output side includes multiple output windows 152. In this embodiment, each output window 152 is connected to multiple input windows 158 via transparent pathways 156 through the reference structure 154, while the remainder of the reference structure 154 body is opaque. A prototype version of reference structure 154 has been constructed using a stereolithographic printer. A photograph of the sensor is shown in FIG. 49.
  • This reference structure has been used to track motion. Data from the sensor is shown in FIGS. 50 and 51. FIG. 50 shows the reference structure signal as a function of time and channel number, and FIG. 51 shows the inverse transform for source position (a car going by) as a function of time. [0191]
  • The foregoing embodiments and advantages are merely exemplary and are not to be construed as limiting the present invention. The present teaching can be readily applied to other types of apparatuses. The description of the present invention is intended to be illustrative, and not to limit the scope of the claims. Many alternatives, modifications, and variations will be apparent to those skilled in the art. In the claims, means-plus-function clauses are intended to cover the structures described herein as performing the recited function and not only structural equivalents but also equivalent structures. [0192]
  • Appendix A
  • [0193] N = orderOfGrid ( NbyNgrid ) p 1 = { [ N / 2 , N / 2 ] Neven [ ( N + 1 ) / 2 , ( N + 1 ) / 2 ] Nodd p 2 = p 1 + [ 0 , 1 ] p 3 = p 1 + [ 1 , 1 ] p 4 = p 1 + [ 1 , 0 ] p EvenSquares = p 1 + [ N / 2 , 1 - N / 2 ] p OddSquares = p 1 - [ ( N - 1 ) / 2 , ( N - 1 ) / 2 ] dvec ( m ) = ( - 1 ) chop ( m ) + 1 { [ 0 , 1 ] m - chop ( m ) 2 = 1 [ m - 1 , 1 ] 1 < m - chop ( m ) 2 chop ( m ) 2 + chop ( m ) + 1 [ chop ( m ) , chop ( m ) ( chop ( m ) + 1 ) - m ] m - chop ( m ) 2 > chop ( m ) 2 + chop ( m ) + 1 p m = p chop ( m ) + dvec ( m )
    Figure US20040095626A1-20040520-M00032

Claims (88)

What is claimed is:
1. A source resolver, comprising:
a multi-dimensional reference structure configured to modulate source radiation in a predetermined manner to produce a reference structure modulated signal comprising a plurality of source space projections, wherein the multi-dimensional reference structure comprises a longitudinal dimension; and
an analyzer configured in accordance with said reference structure for resolving a source state from the reference structure modulated signal.
2. The source resolver of claim 1, wherein the source radiation comprises radiation generated by the source.
3. The source resolver of claim 1, wherein the source radiation comprises radiation scattered by the source.
4. The source resolver of claim 1, wherein the source radiation comprises a radiation reflected by a source.
5. The source resolver of claim 1, wherein the source radiation has propagated through the source.
6. The source resolver of claim 1, wherein the multi-dimensional reference structure modulates the source radiation by attenuating a portion of the source radiation.
7. The source resolver of claim 1, wherein the multi-dimensional reference structure modulates the source radiation by diffracting a portion of the source radiation.
8. The source resolver of claim 1, wherein the multi-dimensional reference structure modulates the source radiation by scattering a portion of the source radiation.
9. The source wave resolver of claim 1, wherein the multi-dimensional reference structure modulates the source radiation by refracting a portion of the source radiation.
10. The source resolver of claim 1, wherein the multi-dimensional reference structure modulates the source radiation by obscuring a portion of the source radiation.
11. The source resolver of claim 10, wherein the multi-dimensional reference structure comprises at least one aperture in a substantially opaque material, wherein the aperture is configured to pass a portion of the source radiation and the substantially opaque material is configured to obscure the source radiation.
12. The source resolver of claim 11, wherein the multi-dimensional reference structure comprises multiple apertures in a substantially opaque material, wherein at least two of the apertures overlap.
13. The source resolver of claim 12, wherein the multiple apertures comprise a first aperture defining a first plane and a second aperture defining a second plane, wherein the first and second planes are different.
14. The source resolver of claim 10, wherein the multi-dimensional reference structure comprises at least one channel in a substantially opaque material, wherein the channel is configured to pass a portion of the source radiation and the substantially opaque material is configured to substantially obscure a portion of the source radiation.
15. The source resolver of claim 14, wherein the at least one channel comprises a first channel defining a first line and a second channel defining a second line, wherein the first and second lines are not parallel.
16. The source resolver of claim 15, wherein the first line intersects the second line.
17. The source resolver of claim 1, wherein the multi-dimensional reference structure modulates the source radiation by absorbing a portion of the source radiation.
18. The source resolver of claim 1, wherein the multi-dimensional reference structure modulates the source radiation by shifting a frequency of a portion of the source radiation.
19. The source resolver of claim 1, wherein the multi-dimensional reference structure modulates the source radiation by diffusing a portion of the source radiation.
20. The source resolver of claim 1, wherein the multi-dimensional reference structure modulates the source radiation by reflecting a portion of the source radiation.
21. The source resolver of claim 20, wherein the multi-dimensional reference structure comprises at least one mirror.
22. The source resolver of claim 21, wherein the multi-dimensional reference structure comprises at least one moveable mirror.
23. The source resolver of claim 22, wherein the multi-dimensional reference structure comprises an array of moveable mirrors.
24. The source resolver of claim 1, wherein the multi-dimensional reference structure comprises at least one sensor.
25. The source resolver of claim 1, wherein the reference structure comprises a structure with transverse and longitudinal dimensions.
26. The source resolver of claim 1, wherein the multi-dimensional reference structure comprises a three-dimensional structure.
27. The source resolver of claim 1, wherein the analyzer comprises a sensor.
28. The source resolver of claim 27, wherein the sensor comprises a plurality of sensors.
29. The source resolver of claim 27, wherein the sensor comprises an array of sensors.
30. The source resolver of claim 29, wherein the array of sensors is substantially periodic.
31. The source resolver of claim 27, wherein the sensor comprises at least one sensor configured to detect at least one reference structure modulated signal parameter.
32. The source resolver of claim 31, wherein the at least one reference structure modulated signal parameter comprises amplitude.
33. The source resolver of claim 31, wherein the at least one reference structure modulated signal parameter comprises phase.
34. The source resolver of claim 31, wherein the at least one reference structure modulated signal parameter comprises frequency.
35. The source resolver of claim 31, wherein the at least one reference structure modulated signal parameter comprises polarity.
36. The source resolver of claim 31, wherein the at least one reference structure modulated signal parameter comprises dispersion.
37. The source resolver of claim 31, wherein the at least one reference structure modulated signal parameter comprises carrier frequency
38. The source resolver of claim 36, wherein the source radiation comprises wave packets and the at least one reference structure modulated signal parameter comprises a wavelength of an envelope of the wave packet.
39. The source resolver of claim 1, wherein the analyzer comprises a processor.
40. The source resolver of claim 1, wherein the multi-dimensional reference structure comprises an active reference structure coupled to a reference structure driver.
41. The source resolver of claim 40, wherein the reference structure driver is in communication with the analyzer.
42. The source resolver of claim 40, wherein the reference structure driver is in communication with the source.
43. The source resolver of claim 41, wherein the reference structure driver is configured to receive a feedback signal from the analyzer and output a control signal to the active reference structure.
44. The source resolver of claim 43, wherein the modulation of the source radiation by the active reference structure is variable in accordance with the control signal from the reference structure driver.
45. A tomography machine configured to image an object, comprising the source resolver of claim 1.
46. The source resolver of claim 1, wherein the plurality of source space projections are produced by the multi-dimensional reference structure in a single time step.
47. The source resolver of claim 1, wherein the plurality of source space projections are produced by the multi-dimensional reference structure without scanning of the multi-dimensional reference structure.
48. A source imager, comprising:
a multi-dimensional reference structure configured to modulate source radiation to produce a reference structure modulated signal comprising a plurality of source space projections, wherein the multi-dimensional reference structure comprises a longitudinal dimension;
a sensor array configured to detect the reference structure modulated signal; and
a processor coupled to the sensor array and configured to derive source information from the detected reference structure modulated signal.
49. The source imager of claim 48, wherein the source information comprises a state of the source.
50. The source imager of claim 48, wherein the source information comprises an image of the source.
51. The source imager of claim 48, wherein the plurality of source space projections are produced by the multi-dimensional reference structure in a single time step.
52. The source imager of claim 48, wherein the plurality of source space projections are produced by the multi-dimensional reference structure without scanning of the multi-dimensional reference structure.
53. The source imager of claim 48, wherein the multi-dimensional reference structure comprises a three-dimensional reference structure.
54. A tomographic apparatus, comprising:
a source;
a sensor;
a multi-dimensional reference structure between the source and the sensor for producing modulated radiation from source radiation, wherein the modulated radiation is received by the sensor and comprises a plurality of source space projections, and wherein the multi-dimensional reference structure comprises a longitudinal dimension; and
an analyzer coupled to the sensor for analyzing the modulated radiation.
55. The apparatus of claim 54, wherein the plurality of source space projections are produced by the multi-dimensional reference structure in a single time step.
56. The apparatus of claim 54, wherein the plurality of source space projections are produced by the multi-dimensional reference structure without scanning of the multi-dimensional reference structure.
57. The apparatus of claim 54, wherein the multi-dimensional reference structure comprises a three-dimensional reference structure.
58. A method of enhancing a sensor, comprising:
altering radiation received by the sensor by inserting a multi-dimensional reference structure into a path of the radiation, wherein the altered radiation comprises a plurality of source space projections, and wherein the multi-dimensional reference structure comprises a longitudinal dimension; and
analyzing the sensor's response to the altered radiation.
59. The method of claim 58, wherein the plurality of source space projections are produced by the multi-dimensional reference structure in a single time step.
60. The apparatus of claim 58, wherein the plurality of source space projections are produced by the multi-dimensional reference structure without scanning of the multi-dimensional reference structure.
61. The method of claim 58, wherein the multi-dimensional reference structure comprises a three-dimensional reference structure.
62. A sensor system, comprising:
a multi-dimensional reference structure configured to receive unconditioned data radiated from a source and to produce conditioned data comprising a plurality of source space projections, wherein the multi-dimensional reference structure comprises a longitudinal dimension;
a sensor configured to receive the conditioned data; and
an analyzer configured to invert the conditioned data received by the sensor.
63. The sensor system of claim 62, wherein producing conditioned data comprises creating a linear transformation between the unconditioned data and the conditioned data with the multi-dimensional reference structure.
64. The sensor system of claim 63, wherein the reference structure is configured to create a well-conditioned linear transformation.
65. The sensor system of claim 62, wherein the unconditioned data comprises space-time coding.
66. The sensor system of claim 62, wherein the multi-dimensional reference structure comprises a non-monotonic multi-dimensional object.
67. The sensor system of claim 62, wherein the plurality of source space projections are produced by the multi-dimensional reference structure in a single time step.
68. The sensor system of claim 62, wherein the plurality of source space projections are produced by the multi-dimensional reference structure without scanning of the multi-dimensional reference structure.
69. The sensor system of claim 58, wherein the multi-dimensional reference structure comprises a three-dimensional reference structure.
70. A measurement system, comprising:
a source configured to produce source modulated radiation;
a multi-dimensional reference structure comprising a longitudinal dimension for receiving the source modulated radiation and producing reference modulated wave radiation comprising a plurality of source space projections; and
an analyzer in communication with the reference modulated radiation for analyzing the reference modulated radiation.
71. The system of claim 70, wherein the source modulated radiation is time modulated.
72. The system of claim 70, wherein the source modulated radiation is space modulated.
73. The system of claim 70, wherein the source modulated radiation is space-time modulated.
74. The system of claim 70, wherein the source modulated radiation comprises a controllably variable pattern.
75. The system of claim 70, wherein the source comprises a controllably variable filter.
76. The system of claim 75, wherein the controllably variable filter comprises an LCD (liquid crystal display).
77. The system of claim 70, wherein the plurality of source space projections are produced by the multi-dimensional reference structure in a single time step.
78. The system of claim 70, wherein the plurality of source space projections are produced by the multi-dimensional reference structure without scanning of the multi-dimensional reference structure.
79. The system of claim 70, wherein the multi-dimensional reference structure comprises a three-dimensional reference structure.
80. A method of resolving a source state, comprising:
modulating source radiation such that a plurality of source space projections are created in a parallel fashion; and
resolving the source state using the plurality of source space projections.
81. The method of claim 80, wherein resolving the source state comprises reconstructing a source distribution.
82. The method of claim 80, wherein resolving the source state comprises determining a spatial position of the source.
83. The method of claim 80, wherein the source radiation is modulated along transverse and longitudinal dimensions.
84. The method of claim 83, wherein the source radiation is modulated with a multi-dimensional reference structure.
85. The method of claim 83, wherein the source radiation is modulated with a volumetric reference structure.
86. The method of claim 80, wherein the source radiation comprises electromagnetic radiation.
87. The method of claim 80, wherein the source radiation comprises acoustic radiation.
88. The method of claim 80, wherein the source radiation comprises optical radiation.
US10/388,480 2002-09-30 2003-03-17 Reference structures and reference structure enhanced tomography Abandoned US20040095626A1 (en)

Priority Applications (4)

Application Number Priority Date Filing Date Title
US10/388,480 US20040095626A1 (en) 2002-09-30 2003-03-17 Reference structures and reference structure enhanced tomography
PCT/US2003/011390 WO2004032047A2 (en) 2002-09-30 2003-04-14 Reference structures and reference structure enhanced tomography
AU2003224956A AU2003224956A1 (en) 2002-09-30 2003-04-14 Reference structures and reference structure enhanced tomography
US12/000,450 US7912173B2 (en) 2002-09-30 2007-12-12 Reference structures and reference structure enhanced tomography

Applications Claiming Priority (3)

Application Number Priority Date Filing Date Title
US41434402P 2002-09-30 2002-09-30
US41436202P 2002-09-30 2002-09-30
US10/388,480 US20040095626A1 (en) 2002-09-30 2003-03-17 Reference structures and reference structure enhanced tomography

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US12/000,450 Continuation US7912173B2 (en) 2002-09-30 2007-12-12 Reference structures and reference structure enhanced tomography

Publications (1)

Publication Number Publication Date
US20040095626A1 true US20040095626A1 (en) 2004-05-20

Family

ID=32074330

Family Applications (2)

Application Number Title Priority Date Filing Date
US10/388,480 Abandoned US20040095626A1 (en) 2002-09-30 2003-03-17 Reference structures and reference structure enhanced tomography
US12/000,450 Expired - Fee Related US7912173B2 (en) 2002-09-30 2007-12-12 Reference structures and reference structure enhanced tomography

Family Applications After (1)

Application Number Title Priority Date Filing Date
US12/000,450 Expired - Fee Related US7912173B2 (en) 2002-09-30 2007-12-12 Reference structures and reference structure enhanced tomography

Country Status (3)

Country Link
US (2) US20040095626A1 (en)
AU (1) AU2003224956A1 (en)
WO (1) WO2004032047A2 (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20050058352A1 (en) * 2003-07-16 2005-03-17 Shrenik Deliwala Optical encoding and reconstruction
US20050117701A1 (en) * 2003-12-01 2005-06-02 Nelson James M. Backscatter imaging using hadamard transform masking
US20180010911A1 (en) * 2016-04-19 2018-01-11 Massachusetts Institute Of Technology Ground-Based System for Geolocation of Perpetrators of Aircraft Laser Strikes

Families Citing this family (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8000435B2 (en) 2006-06-22 2011-08-16 Koninklijke Philips Electronics N.V. Method and system for error compensation
WO2013103408A1 (en) 2011-10-07 2013-07-11 Duke University Apparatus for coded aperture x-ray scatter imaging and method therefor
JP5875953B2 (en) * 2012-07-17 2016-03-02 株式会社東芝 Optical device
US20140104979A1 (en) * 2012-08-09 2014-04-17 Stolar Research Corporation Ground-Penetrating Tunnel-Detecting Active Sonar
US10004464B2 (en) 2013-01-31 2018-06-26 Duke University System for improved compressive tomography and method therefor
US10107768B2 (en) 2013-08-13 2018-10-23 Duke University Volumetric-molecular-imaging system and method therefor
US10559085B2 (en) * 2016-12-12 2020-02-11 Canon Kabushiki Kaisha Devices, systems, and methods for reconstructing the three-dimensional shapes of objects
US10987071B2 (en) * 2017-06-29 2021-04-27 University Of Delaware Pixelated K-edge coded aperture system for compressive spectral X-ray imaging
US10909670B2 (en) 2018-12-06 2021-02-02 Massachusetts Institute Of Technology Computational reconfigurable imaging spectrometer

Citations (15)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US4075483A (en) * 1976-07-12 1978-02-21 Raytheon Company Multiple masking imaging system
US4165462A (en) * 1977-05-05 1979-08-21 Albert Macovski Variable code gamma ray imaging system
US4506374A (en) * 1982-04-08 1985-03-19 Technicare Corporation Hybrid collimator
US5036546A (en) * 1986-10-06 1991-07-30 Grumman Aerospace Corporation Pseudo-noise product coded aperture arrays and method for designing same
US5245191A (en) * 1992-04-14 1993-09-14 The Board Of Regents Of The University Of Arizona Semiconductor sensor for gamma-ray tomographic imaging system
US5606165A (en) * 1993-11-19 1997-02-25 Ail Systems Inc. Square anti-symmetric uniformly redundant array coded aperture imaging system
US5627639A (en) * 1995-06-06 1997-05-06 Lockheed Missiles & Space Company, Inc. Coded aperture imaging spectrometer
US5772598A (en) * 1993-12-15 1998-06-30 Forschungszentrum Julich Gmbh Device for transillumination
US5930314A (en) * 1996-05-31 1999-07-27 Massachusetts Institute Of Technology Coded aperture imaging
US6046808A (en) * 1999-04-09 2000-04-04 Three Lc, Inc. Radiation filter, spectrometer and imager using a micro-mirror array
US20010056234A1 (en) * 2000-04-12 2001-12-27 Weinberg Irving N. Hand held camera with tomographic capability
US6392235B1 (en) * 1999-02-22 2002-05-21 The Arizona Board Of Regents On Behalf Of The University Of Arizona Coded-aperture system for planar imaging of volumetric sources
US20020075990A1 (en) * 2000-09-29 2002-06-20 Massachusetts Institute Of Technology Coded aperture imaging
US6583420B1 (en) * 2000-06-07 2003-06-24 Robert S. Nelson Device and system for improved imaging in nuclear medicine and mammography
US20060108509A1 (en) * 2004-09-09 2006-05-25 Frangioni John V Systems and methods for multi-modal imaging

Family Cites Families (18)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US3746872A (en) * 1971-07-27 1973-07-17 Nuclear Chicago Corp Tomography technique in which a single recording film retains spatial information to permit constructing all planar sections of object
US3867637A (en) * 1973-09-04 1975-02-18 Raytheon Co Extended monochromatic x-ray source
DE2514988C3 (en) * 1975-04-05 1980-03-13 Philips Patentverwaltung Gmbh, 2000 Hamburg Method for the layer-by-layer representation of three-dimensional objects by means of coded simultaneous overlay images
DE2616714A1 (en) * 1976-04-15 1977-10-27 Philips Patentverwaltung METHOD FOR LAYERED REPRESENTATION OF OBJECTS FROM OVERLAY IMAGES OF DIFFERENT IMAGE LEVELS
US4209780A (en) * 1978-05-02 1980-06-24 The United States Of America As Represented By The United States Department Of Energy Coded aperture imaging with uniformly redundant arrays
DE2943758C2 (en) * 1979-10-30 1985-07-11 Philips Patentverwaltung Gmbh, 2000 Hamburg Method for spatial decoding of three-dimensional objects from primary overlay images produced by means of multiple beam sources
DE3037621A1 (en) * 1980-10-04 1982-05-27 Philips Patentverwaltung Gmbh, 2000 Hamburg TRANSLUCTION ARRANGEMENT FOR TAKING LAYER IMAGES OF A THREE-DIMENSIONAL OBJECT
DE3121324A1 (en) * 1981-05-29 1982-12-16 Philips Patentverwaltung Gmbh, 2000 Hamburg TOMOSYNTHESIS DEVICE
DE3209683A1 (en) * 1982-03-17 1983-09-22 Philips Patentverwaltung Gmbh, 2000 Hamburg X-RAY TARGET DEVICE WITH A FILM FADING DEVICE
DE3433109A1 (en) * 1984-09-08 1986-03-27 Volker Dipl.-Ing. 2800 Bremen Döring DEVICE AND METHOD FOR DISPLAYING THE DISTRIBUTION OF HIGH ACTIVITIES OF RADIOACTIVE SUBSTANCES WITH THE AID OF WHOLE-BODY SCINTIGRAPHY DEVICES
US4651012A (en) * 1985-03-21 1987-03-17 Martin Marietta Corporation High brilliance lensless projection system of test patterns
JP2606207B2 (en) * 1987-03-25 1997-04-30 株式会社島津製作所 Imaging equipment
DE3909147A1 (en) * 1988-09-22 1990-09-27 Philips Patentverwaltung ARRANGEMENT FOR MEASURING THE IMPULSE TRANSFER
US4998268A (en) * 1989-02-09 1991-03-05 James Winter Apparatus and method for therapeutically irradiating a chosen area using a diagnostic computer tomography scanner
DE4034602A1 (en) * 1990-06-20 1992-05-07 Philips Patentverwaltung ARRANGEMENT FOR MEASURING THE IMPULSE TRANSMISSION SPECTRUM OF ROENTGEN QUANTS
FR2699319B1 (en) * 1992-12-14 1995-03-10 Von Laue Paul Langevin Inst Ma Device for collimating beams of radiation.
US5717733A (en) * 1995-05-31 1998-02-10 Quanta Vision, Inc. X-ray and neutron diffractometric imaging of the internal structure of objects
US7375338B1 (en) * 2007-03-07 2008-05-20 General Electric Company Swappable collimators method and system

Patent Citations (18)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US4075483A (en) * 1976-07-12 1978-02-21 Raytheon Company Multiple masking imaging system
US4165462A (en) * 1977-05-05 1979-08-21 Albert Macovski Variable code gamma ray imaging system
US4506374A (en) * 1982-04-08 1985-03-19 Technicare Corporation Hybrid collimator
US5036546A (en) * 1986-10-06 1991-07-30 Grumman Aerospace Corporation Pseudo-noise product coded aperture arrays and method for designing same
US5245191A (en) * 1992-04-14 1993-09-14 The Board Of Regents Of The University Of Arizona Semiconductor sensor for gamma-ray tomographic imaging system
US5606165A (en) * 1993-11-19 1997-02-25 Ail Systems Inc. Square anti-symmetric uniformly redundant array coded aperture imaging system
US5772598A (en) * 1993-12-15 1998-06-30 Forschungszentrum Julich Gmbh Device for transillumination
US5627639A (en) * 1995-06-06 1997-05-06 Lockheed Missiles & Space Company, Inc. Coded aperture imaging spectrometer
US5930314A (en) * 1996-05-31 1999-07-27 Massachusetts Institute Of Technology Coded aperture imaging
US6205195B1 (en) * 1996-05-31 2001-03-20 Massachusetts Institute Of Technology Coded aperture imaging
US6392235B1 (en) * 1999-02-22 2002-05-21 The Arizona Board Of Regents On Behalf Of The University Of Arizona Coded-aperture system for planar imaging of volumetric sources
US6046808A (en) * 1999-04-09 2000-04-04 Three Lc, Inc. Radiation filter, spectrometer and imager using a micro-mirror array
US20010056234A1 (en) * 2000-04-12 2001-12-27 Weinberg Irving N. Hand held camera with tomographic capability
US6583420B1 (en) * 2000-06-07 2003-06-24 Robert S. Nelson Device and system for improved imaging in nuclear medicine and mammography
US20030205676A1 (en) * 2000-06-07 2003-11-06 Nelson Robert Sigurd Device and system for improved imaging in nuclear medicine and mamMography
US20020075990A1 (en) * 2000-09-29 2002-06-20 Massachusetts Institute Of Technology Coded aperture imaging
US6737652B2 (en) * 2000-09-29 2004-05-18 Massachusetts Institute Of Technology Coded aperture imaging
US20060108509A1 (en) * 2004-09-09 2006-05-25 Frangioni John V Systems and methods for multi-modal imaging

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20050058352A1 (en) * 2003-07-16 2005-03-17 Shrenik Deliwala Optical encoding and reconstruction
US7339170B2 (en) * 2003-07-16 2008-03-04 Shrenik Deliwala Optical encoding and reconstruction
US20050117701A1 (en) * 2003-12-01 2005-06-02 Nelson James M. Backscatter imaging using hadamard transform masking
US6950495B2 (en) * 2003-12-01 2005-09-27 The Boeing Company Backscatter imaging using Hadamard transform masking
US20180010911A1 (en) * 2016-04-19 2018-01-11 Massachusetts Institute Of Technology Ground-Based System for Geolocation of Perpetrators of Aircraft Laser Strikes
US10718613B2 (en) * 2016-04-19 2020-07-21 Massachusetts Institute Of Technology Ground-based system for geolocation of perpetrators of aircraft laser strikes

Also Published As

Publication number Publication date
AU2003224956A8 (en) 2004-04-23
US7912173B2 (en) 2011-03-22
US20100322372A1 (en) 2010-12-23
AU2003224956A1 (en) 2004-04-23
WO2004032047A3 (en) 2007-09-13
WO2004032047A2 (en) 2004-04-15

Similar Documents

Publication Publication Date Title
US7912173B2 (en) Reference structures and reference structure enhanced tomography
Brunelli Template matching techniques in computer vision: theory and practice
Easton Jr Fourier methods in imaging
CA2504620C (en) Method and apparatus for three-dimensional imaging in the fourier domain
Brady et al. Compressive tomography
US7840053B2 (en) System and methods for tomography image reconstruction
US9232211B2 (en) System and methods for three-dimensional imaging of objects in a scattering medium
US10107768B2 (en) Volumetric-molecular-imaging system and method therefor
CN107209362A (en) Fourier overlapping associations tomography
US8294712B2 (en) Scalable method for rapidly detecting potential ground vehicle under cover using visualization of total occlusion footprint in point cloud population
Florescu et al. Single-scattering optical tomography
Lin et al. Passive non-line-of-sight imaging using plenoptic information
IL172186A (en) Method for fast image reconstruction with compact radiation source and detector arrangement using computerized tomography
US20130256518A1 (en) Synthetic focal plane imager
US6810141B2 (en) Method for processing spatial-phase characteristics of electromagnetic energy and information conveyed therein
Witten et al. Geophysical imaging with arbitrary source illumination
Jaffe A tomographic approach to inverse Mie particle characterization from scattered light
Schmalz Errors inherent in the restoration of imagery acquired through remotely sensed refractive interfaces and scattering media
Lindell Computational Imaging with Single-Photon Detectors
Lindström A method for estimating the fractal dimension of digital color images
Levoy Computational imaging in the sciences
Epstein Mathematics of medical imaging
Løvås et al. Semi-analytical model for deep-water hyperspectral imaging
Garces et al. Tomographic Fourier Telescopy
Johnson Passive ranging systems using orthogonal encoding

Legal Events

Date Code Title Description
AS Assignment

Owner name: UNIVERSITY, DUKE, NORTH CAROLINA

Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNOR:BRADY, DAVID J.;REEL/FRAME:013885/0942

Effective date: 20030313

STCB Information on status: application discontinuation

Free format text: ABANDONED -- FAILURE TO RESPOND TO AN OFFICE ACTION