CN101551907B - Method for multi-camera automated high-precision calibration - Google Patents
Method for multi-camera automated high-precision calibration Download PDFInfo
- Publication number
- CN101551907B CN101551907B CN2009100979957A CN200910097995A CN101551907B CN 101551907 B CN101551907 B CN 101551907B CN 2009100979957 A CN2009100979957 A CN 2009100979957A CN 200910097995 A CN200910097995 A CN 200910097995A CN 101551907 B CN101551907 B CN 101551907B
- Authority
- CN
- China
- Prior art keywords
- calibration
- unique point
- picture
- unique
- point
- 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.)
- Expired - Fee Related
Links
Images
Abstract
The invention discloses a method for multi-camera automated high-precision calibration, including following steps: 1, setting calibration object; 2, screening calibration pictures; 3, extracting feature pointf of the screened calibration pictures; 4, matching the feature points; 5, putting information of the matched feature points into W matrix to repair the missing W matrix; 6, using the Tomas svoboda algorithm to decompose the repaired W matrix for calculating internal and external parameter of the camers. The invention provides a method for multi-camera automated high-precision calibration with simple calibration object set, which can calibrating automatically with high-precision and high-speed.
Description
Technical field
The present invention relates to calibration technology, particularly method for multi-camera automated high-precision calibration.
Background technology
Camera Calibration mainly is in order to determine the inner parameter and the external parameter of camera.Inner parameter mainly comprises picture centre (u0 and v0), focal distance f (ax as a result and ay after also the product of it and magnification being lumped together sometimes call focal length), inclination coefficient s, distortion parameter etc.When external parameter is meant and takes camera in world coordinate system the position and towards, can describe with a rotation matrix R and a translation vector T.When not considering distortion parameter, available following formula is represented:
[X a bit in the world coordinate system wherein
w, Y
w, Z
w]
TImage space in photo is [u, v] T.For the distortion parameter of camera lens, mainly be in proposition plummet model (Plumb Bobmodel) in 1966 with Brown.Concrete grammar is as follows: postulated point P is [X in camera coordinates system
c, Y
c, Z
c]
T, with the coordinate [X of P point in camera coordinates system
c, Y
c, Z
c]
TThe projection coordinate that all obtains a P after the normalization is [X
c/ Z
c, Y
c/ Z
c]
T=[x, y]
TMake r
2=x
2+ y
2, then after lens distortion calibration, new normalization coordinate is [x+x
r+ x
t, y+y
r+ y
t]
T, [x wherein
r, y
r]
TBe radial distortion vector, [x
r, y
r]
T=[kc (1) r
2+ kc (2) r
4+ kc (5) r
6] [x, y]
T[x
t, y
t]
TBe tangential distortion vector, x
t=2kc (3) xy+kc (4) (r
2+ 2x
2), y
t=2kc (4) xy+kc (3) (r
2+ 2y
2); After calibration, just determined the correspondent transform relation between camera coordinates system and the image coordinate system, thereby can be used for object identification, estimation, three-dimensional body reconstruction, texture etc.
As shown in Figure 1, the situation and the one camera of polyphaser calibration are similar, some x
jIn camera C
1, C
2, C
3, C
4Middle imaging is u
j 1, u
j 2, u
j 3, u
j 4, for the parameter matrix p of each camera
1, p
2, p
3, p
4, according to the top model that we are set forth, we can access:
So, for a plurality of three-dimensional point, x
1X
n, we have following equation:
W
s＝PX
Existing multi-camera scaling algorithm is divided and can be divided into according to the calibration thing: do not have the calibration thing, 1 dimension drift slide, 2 dimension calibration plates and three-dimensional calibration body.
The method of demarcating camera based on 0 dimension does not only rely on any demarcation thing to come the calibrating camera intrinsic parameter by the correspondence of the point between image.These methods all are to utilize the picture of absolute conic or its antithesis absolute quadric only relevant with the intrinsic parameter of video camera and demarcate with the irrelevant character of camera motion.The advantage of this method is more flexible than additive method, need not to calibrate the support of thing, yet its shortcoming is obvious too: algorithm all is non-linear, needs very complicated calculating and calibration result robust not usually.
Calibrating method based on 2 dimension calibration plates is the multiple image of different azimuth camera plane template or drift slide, utilizes the spatial point of plane template and the correspondence between picture point to set up equation of constraint realization Camera calibration.The advantage of this method is the 3 dimension information that do not need prior known scaled plate, calibration precision is higher, but these class methods have that a fatal shortcoming-when the calibration plate did not have complete the time to be caught by all cameras, the result that scaling algorithm can't be carried out or calibrate was extremely unstable.
Scaling method based on three-dimensional calibration body requires video camera that the accurate known demarcation object of three-dimensional geometric information is taken, and sets up equation of constraint according to the three-dimensional point of demarcating object and the correspondence of picture point, carries out Camera calibration.Though this method calibration precision is higher, calibrates object and involves great expense.
At present, popular in the world method is the method based on 1 dimension scaling point that Tomas Svoboda proposes.Extraction for scaling point is usually found the led light source center with bicubic interpolation and 2D Gaussian filtering.Because we are to the improvement of calibration thing, need the result to be divided into 4 classes before the extraction for scaling point with the K-Means method.The right precision of the method for Tomas is high and full-automatic, but the calibration time is very slow.
Summary of the invention
The invention provides a kind of calibration thing and be provided with simply, calibration precision height, fireballing method for multi-camera automated high-precision calibration.
A kind of method for multi-camera automated high-precision calibration may further comprise the steps:
(1) the calibration thing is set;
Described calibration thing is by a, and the led light source that b, c, four of d have a proportionate relationship is fixed on the plank to be formed; Wherein a, b, three led light source conllinear of c, a, b, three led light source places of c straight line and b, two led light source places of d straight line are vertical, and the distance of a and b is to be a unit length, and the distance of b and d is 2 unit lengths, and the distance of b and c is 4 unit lengths; Can be when utilization according to any selection unit of the size length of scene.
In scene is under 10 meters square situations, and for obtaining to calibrate the result preferably, unit length can be selected 20cm.It is the coupling of back unique point for convenience that four led light sources are designed to 1: 2: 4 proportionate relationship purpose.
(2) take the calibration picture;
Settle a plurality of parameters identical camera and the good camera lens of manual adjustment in scene after, the hand-held calibration of operator thing moves in whole scene, and a plurality of cameras are taken several calibration pictures to the calibration thing simultaneously, takes 100 calibration pictures for one group.
(3) extract unique point on the captured calibration picture;
1) calculate the average image of captured calibration picture: every value of calibrating respective pixel in the picture that will be captured adds up, and then divided by the number of captured calibration picture, obtain the mean value of each pixel on the captured calibration picture, thereby obtain the average image of captured calibration picture, will reduce the influence of surround lighting in the binaryzation like this.
2) each pixel value in the calibration picture that will take deducts in the average image corresponding pixel value and writes down maximum difference; With pixel value difference greater than 90% pixel of maximum difference as candidate point, then will the pixel in candidate point continue to deduct in the average image corresponding pixel value and write down maximum difference, with pixel value difference greater than 90% pixel of maximum difference as candidate point, carry out the cluster analysis screening with the K-MEANS method again and obtain four groups of collection of pixels, one group of collection of pixels is a led light source.
3) remove undesirable collection of pixels;
The size of the led light source that is provided with greater than us when the number of led light source set and collection of pixels are undesirable collection of pixels when not being communicated with, with its removal; The size of described led light source is set to 10 pixels usually;
4) utilize bicubic interpolation and 2D Gaussian filtering to find each collection of pixels, i.e. the sub-pix center of circle of each led light source, the described sub-pix center of circle are the unique point on the calibration picture.
(4) unique point is mated;
Special coupling on schedule is the critical process that calibration is calculated, and each camera of synchronization finds many more match points between taking a picture, and result calculated is accurate more, and efficient is high more.
I) unique point on the calibration picture is (promptly to calibrate thing and take situation completely) under four situations, carries out following steps:
(1) each unique point in four unique points is sorted apart from sum to other 3, find out a unique point apart from the sum minimum;
(2) will carry out ascending ordering to the distance of other three unique points apart from the unique point of sum minimum, judge whether the ratio of three distances after sorting satisfied 1: 2: 4, if satisfied four unique points of then being looked for are for calibrating the point of thing; Otherwise mistake appears in the extraction of the unique point on the calibration picture, and we cast out this group photo.
II) under the situation of the unique point d disappearance on the calibration picture, carry out following steps:
Whether the slope that calculates line between three unique points equates, if equate described three the unique point conllinear of explanation, judge have minor increment and unique point whether satisfy 1: 4 relation to other distances of 2, if satisfy, then determine the matching relationship of a, b, three unique points of c, error gets final product less than 0.01 in actual applications.
If do not satisfy, can carry out polishing by replenishing the W matrix for the unique point d point that lacks.
III) under the situation of the unique point b disappearance on the calibration picture, carry out following steps:
Whether satisfied 1: 2 after judging 3 line ascending sort: sqrt (5), can determine that then be unique point a, c, d at these 3 if satisfy, error gets final product less than 0.01 in actual applications.
If do not satisfy, can carry out polishing by replenishing the W matrix for the unique point b point that lacks;
IV) situation of the unique point disappearance on the calibration picture is except that above-mentioned three kinds of situations, and the unique point on the calibration picture can't be mated; If the unique point in one group of calibration picture that camera is taken on all pictures all can't be mated, then ignore this group calibration picture, next group calibration picture is carried out Feature Points Matching; Otherwise remain picture if the unique point in some the calibration picture in one group of calibration picture that camera is taken can't be mated passable, though can not determine the relation between the unique point, can carry out polishing by replenishing the W matrix.
The match point that finds between each camera picture of synchronization is many more, and result calculated is accurate more, and efficient is high more, 4 Feature Points Matching of the calibration thing in the calibration picture that each camera is at a time taken promptly as much as possible.
(5) information of the unique point after will mating is put into the W matrix, carries out polishing for the W matrix of disappearance;
Calculate its projection degree of depth according to existing unique point, be that 4 W ' matrix is according to matrix W according to orders of these information architectures such as the projection degree of depth then ' disappearance information among the filled matrix W, specifically the disappearance point polishing method of the W matrix that proposes of Daniel Martinec that mentions with reference to " Fill of missing elements in " chapters and sections in the paper (list of references 1) " A Convenient Multi-CameraSelf-Calibration for Virtual Environments " of Tomas svoboda and Tomas Pajdla is carried out the polishing of W matrix.List of references 1:Tomas Svoboda, Daniel Martinec, andTomas Pajdla. " A convenient multi-camera self-calibraion for virtualenvironments " .PRESENCE:Teleoperators and Virtual Environments, pp407-422,14 (4), August 2005.MIT PRESS.
(6) algorithm with the utilization of the W matrix behind polishing Tomas svoboda carries out split-matrix, calculates the camera inside and outside parameter.Specifically with reference to the method in " Euclideanstratification " part in the paper " A ConvenientMulti-Camera Self-Calibration for Virtual Environments " of Tomas svoboda.
A kind of method for multi-camera automated high-precision calibration of the present invention has the following advantages:
1. algorithm robust.Find the solution the camera inside and outside parameter in the algorithm and adopted ripe in the world decomposition algorithm.And, be different from conventional two-dimensional calibration plate algorithm, when the calibration thing was not obtained by all cameras fully, new algorithm still can be discerned available point or fill up the disappearance point, thereby guaranteed the robustness of algorithm.
2. auto-scaling.The just hand-held calibration thing of all operation of calibrator moves in scene, wait to have taken calibrate picture after, need not artificial participation and can calculate the calibration result.
3. calibration at a high speed.Because the available point in the original calibration thing had become four by one, so improved the speed of calibration greatly, under the situation with new-type calibration thing, the number of calibrating photograph taking significantly reduces, and is equivalent to original 1/4th.And, under processing speed and original roughly the same situation, obtained than the more point of original algorithm to every photo.Obtain the original minimizing 3/4ths of time ratio of the point of identical scale.Obtaining n available point in the old algorithm needs n to open image at least, and the inventive method only needs n/4 to open the calibration picture.
Description of drawings
Fig. 1 is the synoptic diagram of polyphaser calibration;
Fig. 2 is the process flow diagram of method for multi-camera automated high-precision calibration of the present invention;
Fig. 3 is the structural representation of four calibrations of the present invention thing;
Fig. 4 is that four of the present invention calibration thing is taken the synoptic diagram when complete;
Fig. 5 is that four calibrations of the present invention thing is taken the synoptic diagram that lacks one of them
Fig. 6 is that four calibrations of the present invention thing is taken the synoptic diagram that lacks another.
Embodiment
A kind of method for multi-camera automated high-precision calibration of the present invention as shown in Figure 2, may further comprise the steps: 1. the calibration thing is set; 2. take the calibration picture; 3. extract the unique point on the captured calibration picture; 4. unique point is mated; 5. the information of the unique point after will mating is put into the W matrix, carries out polishing for the W matrix that lacks; 6. the algorithm with the utilization of the W matrix behind polishing Tomas svoboda carries out split-matrix, calculates the camera inside and outside parameter.
(1) the calibration thing is set;
As shown in Figure 3, the calibration thing is by a, b, the led light source that c, four of d have a proportionate relationship is fixed on the plank to be formed, wherein a, b, three led light source conllinear of c, a, b, three led light source places of c straight line and b, two led light source places of d straight line are vertical, the distance of a and b is to be a unit length, and the distance of b and d is 2 unit lengths, and the distance of b and c is 4 unit lengths; The scene at place is 10 square metres, and unit length is 20cm.
(2) take the calibration picture;
Settle a plurality of cameras and the good camera lens of manual adjustment in scene after, the hand-held calibration of operator thing moves in whole scene, and a plurality of cameras are taken 100 calibration pictures to the calibration thing simultaneously.
(3) extract unique point on the captured calibration picture;
I) calculate the average image of captured calibration picture: every value of calibrating respective pixel in the picture that will be captured adds up, and then divided by the number of captured calibration picture, obtain the mean value of each pixel on the captured calibration picture, thereby obtain the average image of captured calibration picture.
II) each pixel value in the calibration picture that will take deducts in the average image corresponding pixel value and notes maximum difference; With pixel value difference greater than 90% pixel of maximum difference as the candidate, screening obtains four groups of collection of pixels at the most, one group of collection of pixels is a led light source;
III) remove undesirable collection of pixels;
The size of the led light source that is provided with greater than us when the number of led light source set and collection of pixels are undesirable collection of pixels when not being communicated with, with its removal;
IV) utilize bicubic interpolation and 2D Gaussian to find each collection of pixels, i.e. the sub-pix center of circle of each led light source, the described sub-pix center of circle are the unique point on the calibration picture.
(4) unique point is mated;
Special coupling on schedule is the critical process that calibration is calculated, and each camera of synchronization is afraid of to find between the photo many more match points, and result calculated is accurate more, and efficient is high more.
I) unique point on the calibration picture is under four situations, and (promptly calibrate thing and take situation completely, two kinds of situations of A, B are the calibration thing and take situation completely among Fig. 4) as shown in Figure 4 carries out following steps:
1) each unique point in four unique points is sorted apart from sum to other 3, find out a unique point apart from the sum minimum;
2) will carry out ascending ordering to the distance of other three unique points apart from the unique point of sum minimum, judge whether the ratio of three distances after sorting satisfied 1: 2: 4, if satisfied four unique points of then being looked for are for calibrating the point of thing; Otherwise mistake appears in the extraction of the unique point on the calibration picture, and we cast out this group photo.
II) under the situation of the unique point d disappearance on the calibration picture, as shown in Figure 5, carry out following steps:
Whether the slope that calculates line between three unique points equates, if equate described three the unique point conllinear of explanation, judge have minor increment and unique point whether satisfy 1: 4 relation to other distances of 2, if satisfy, then determine the matching relationship of a, b, three unique points of c;
If do not satisfy, can carry out polishing by replenishing the W matrix for the unique point d point that lacks;
III) under the situation of the unique point b disappearance on the calibration picture, as shown in Figure 6, carry out following steps:
1) whether satisfied 1: 2 after judging 3 line ascending sort: sqrt (5), can determine that then be unique point a, c, d at these 3 if satisfy;
2), can carry out polishing by replenishing the W matrix for the unique point b point that lacks if do not satisfy;
IV) situation of the unique point disappearance on the calibration picture is except that above-mentioned three kinds of situations, and the unique point on the calibration picture can't be mated; If the unique point in one group of calibration picture that camera is taken on all pictures all can't be mated, then ignore this group calibration picture, next group calibration picture is carried out Feature Points Matching; Otherwise remain picture if the unique point in some the calibration picture in one group of calibration picture that camera is taken can't be mated passable, though can not determine the relation between the unique point, can carry out polishing by replenishing the W matrix.
It is many more that each camera of synchronization is clapped the match point that finds between the picture, and result calculated is accurate more, and efficient is high more, 4 Feature Points Matching of the calibration thing in the calibration picture that each camera is at a time taken promptly as much as possible.
(5) information of the unique point after will mating is put into the W matrix, carries out polishing for the W matrix of disappearance;
Calculate its projection degree of depth according to existing unique point, be that 4 W ' matrix is according to matrix W according to orders of these information architectures such as the projection degree of depth then ' disappearance information among the filled matrix W, specifically the disappearance point polishing method of the W matrix that proposes of Daniel Martinec that mentions with reference to " Fill of missing elements in " chapters and sections in the paper (list of references 1) " A Convenient Multi-CameraSelf-Calibration for Virtual Environments " of Tomas svoboda and Tomas Pajdla is carried out the polishing of W matrix.
(6) algorithm with the utilization of the W matrix behind polishing Tomas svoboda carries out split-matrix, calculates the camera inside and outside parameter.Specifically with reference to the method in " Euclideanstratification " part in the paper " A ConvenientMulti-Camera Self-Calibration for Virtual Environments " of Tomas svoboda.
The method of computing camera inside and outside parameter: Tomas svoboda provides the Matlab tool box.At first in matlab, switch to place, tool box catalogue, import the gocal instruction then and can finish demarcation automatically.
Claims (1)
1. a method for multi-camera automated high-precision calibration is characterized in that, may further comprise the steps:
(1) the calibration thing is set; Described calibration thing is by a, b, and c, four led light sources of d are fixed on the plank to be formed; Wherein a, b, three led light source conllinear of c, and a and c are non-conterminous, and a, b, three led light source places of c straight line and b, two led light source places of d straight line are vertical, and the distance of a and b is a unit length, the distance of b and d is 2 unit lengths, and the distance of b and c is 4 unit lengths; Can be when utilization according to any selection unit of the size length of scene;
(2) take the calibration picture;
(3) extract unique point on the captured calibration picture; The extracting method of the unique point on the described calibration picture may further comprise the steps:
(a) calculate the average image of captured calibration picture: every value of calibrating respective pixel in the picture that will be captured adds up, and then divided by the number of captured calibration picture, obtain the mean value of each pixel on the captured calibration picture, thereby obtain the average image of captured calibration picture;
(b) each pixel value in the calibration picture that will take deducts corresponding pixel value and record maximum difference wherein in the average image; Greater than the pixel in 90% the described calibration picture of maximum difference, screening obtains four groups of collection of pixels at the most with pixel value difference, and one group of collection of pixels is a led light source;
(c) remove undesirable collection of pixels;
The size of the LED that is provided with greater than us when the number of led light source set and collection of pixels are undesirable collection of pixels when not being communicated with, with its removal;
(d) utilize bicubic interpolation and 2D Gaussian filtering to find the center point coordinate of each collection of pixels, i.e. the sub-pix center of circle of each led light source, the described sub-pix center of circle are the unique point on the calibration picture;
(4) unique point is mated; The described method that unique point is mated may further comprise the steps:
I) be under four situations in the unique point of calibration on the picture, promptly calibrate thing and take completely under the situation, carry out following steps:
1. each unique point in four unique points is sorted apart from sum to other 3, find out a unique point apart from the sum minimum;
2. will carry out ascending ordering to the distance of other three unique points apart from the unique point of sum minimum, judge whether the ratio of three distances after sorting satisfied 1: 2: 4, if satisfied four unique points of then being looked for are for calibrating the point of thing; Otherwise mistake appears in the extraction of the unique point on the calibration picture, and we cast out this group photo not do calculating;
II) under the situation of the unique point d disappearance on the calibration picture, carry out following steps:
Whether the slope that calculates line between three unique points equates, if equate described three the unique point conllinear of explanation; Each unique point in three unique points is sorted apart from sum to other 2, find out a unique point apart from the sum minimum, judge have minor increment and unique point whether satisfy 1: 4 relation to other distances of 2, if satisfy, then determine the matching relationship of a, b, three unique points of c; If do not satisfy, then go to step (5), by replenishing the W matrix unique point d point is carried out polishing;
III) under the situation of the unique point b disappearance on the calibration picture, carry out following steps:
Whether satisfied 1: 2 after judging 3 line ascending sort: sqrt (5), can determine that then be unique point a, c, d at these 3 if satisfy; Described 3 line ascending sort is: calculate the distance between per two points, one has three points, thus three distance values are arranged, again with these three distance values according to from small to large rank order;
If do not satisfy, then go to step (5), by replenishing the W matrix unique point b point that lacks is carried out polishing;
IV) situation of the unique point disappearance on the calibration picture is except that above-mentioned two kinds of situations, and the unique point on the calibration picture can't be mated; If the unique point in one group of calibration picture that camera is taken on all pictures all can't be mated, then ignore this group calibration picture, next group calibration picture is carried out Feature Points Matching; Otherwise remain picture if the unique point in some the calibration picture in one group of calibration picture that camera is taken can't be mated passable, then goes to step (5), carries out polishing by replenishing the W matrix;
(5) information of the unique point after will mating is put into the W matrix, carries out polishing for the W matrix of disappearance;
(6) algorithm that the W square behind the polishing is fallen utilization Tomas svoboda carries out split-matrix, calculates the camera inside and outside parameter.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2009100979957A CN101551907B (en) | 2009-04-28 | 2009-04-28 | Method for multi-camera automated high-precision calibration |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2009100979957A CN101551907B (en) | 2009-04-28 | 2009-04-28 | Method for multi-camera automated high-precision calibration |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101551907A CN101551907A (en) | 2009-10-07 |
| CN101551907B true CN101551907B (en) | 2011-05-04 |
Family
ID=41156139
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2009100979957A Expired - Fee Related CN101551907B (en) | 2009-04-28 | 2009-04-28 | Method for multi-camera automated high-precision calibration |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101551907B (en) |
Families Citing this family (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102223484A (en) * | 2011-08-04 | 2011-10-19 | 浙江工商大学 | Method and device for configuring head-end parameter of camera |
| CN102436657A (en) * | 2011-09-27 | 2012-05-02 | 夏东 | Active light depth measurement value modifying method based on application of the internet of things |
| IL216515A (en) | 2011-11-22 | 2015-02-26 | Israel Aerospace Ind Ltd | System and method for processing multi-camera array images |
| CN103175851B (en) * | 2011-12-21 | 2015-01-07 | 北京兆维电子(集团)有限责任公司 | Calibrating tool and calibrating method for multi-camera scanning system |
| CN103593836A (en) * | 2012-08-14 | 2014-02-19 | 无锡维森智能传感技术有限公司 | A Camera parameter calculating method and a method for determining vehicle body posture with cameras |
| CN104933755B (en) * | 2014-03-18 | 2017-11-28 | 华为技术有限公司 | A kind of stationary body method for reconstructing and system |
| CN103996220A (en) * | 2014-05-26 | 2014-08-20 | 江苏大学 | Three-dimensional reconstruction method and system in intelligent transportation |
| CN104008548B (en) * | 2014-06-04 | 2017-04-19 | 无锡维森智能传感技术有限公司 | Feature point extraction method for vehicle-mounted around view system camera parameter calibration |
| CN106327587B (en) * | 2016-11-16 | 2019-06-28 | 北京航空航天大学 | A kind of accurate fusion method of laparoscope video for augmented reality surgical navigational |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5768443A (en) * | 1995-12-19 | 1998-06-16 | Cognex Corporation | Method for coordinating multiple fields of view in multi-camera |
| CN1620153A (en) * | 2004-12-02 | 2005-05-25 | 武汉大学 | Method of implementing non-measuration digital camera calibration using plane control point field |
| CN1648600A (en) * | 2004-12-09 | 2005-08-03 | 武汉大学 | Method for three-dimensional measurement of sheet metal part using single non-measuring digital camera |
| CN101231750A (en) * | 2008-02-21 | 2008-07-30 | 南京航空航天大学 | Calibrating method of binocular three-dimensional measuring system |
-
2009
- 2009-04-28 CN CN2009100979957A patent/CN101551907B/en not_active Expired - Fee Related
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5768443A (en) * | 1995-12-19 | 1998-06-16 | Cognex Corporation | Method for coordinating multiple fields of view in multi-camera |
| CN1620153A (en) * | 2004-12-02 | 2005-05-25 | 武汉大学 | Method of implementing non-measuration digital camera calibration using plane control point field |
| CN1648600A (en) * | 2004-12-09 | 2005-08-03 | 武汉大学 | Method for three-dimensional measurement of sheet metal part using single non-measuring digital camera |
| CN101231750A (en) * | 2008-02-21 | 2008-07-30 | 南京航空航天大学 | Calibrating method of binocular three-dimensional measuring system |
Non-Patent Citations (1)
| Title |
|---|
| JP特开2004-135209A 2004.04.30 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN101551907A (en) | 2009-10-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101551907B (en) | Method for multi-camera automated high-precision calibration | |
| Strecha et al. | On benchmarking camera calibration and multi-view stereo for high resolution imagery | |
| CN110276808B (en) | Method for measuring unevenness of glass plate by combining single camera with two-dimensional code | |
| KR101121034B1 (en) | System and method for obtaining camera parameters from multiple images and computer program products thereof | |
| CN105716542B (en) | A kind of three-dimensional data joining method based on flexible characteristic point | |
| CN102376089B (en) | Target correction method and system | |
| CN101915573B (en) | Positioning measurement method based on key point detection of marker | |
| CN108107462B (en) | RTK and high-speed camera combined traffic sign post attitude monitoring device and method | |
| CN105931234A (en) | Ground three-dimensional laser scanning point cloud and image fusion and registration method | |
| CN111429523B (en) | Remote calibration method in 3D modeling | |
| CN109859272A (en) | A kind of auto-focusing binocular camera scaling method and device | |
| CN103702103B (en) | Based on the grating stereo printing images synthetic method of binocular camera | |
| Lou et al. | A cost-effective automatic 3D reconstruction pipeline for plants using multi-view images | |
| Bleier et al. | Low-cost 3d laser scanning in air orwater using self-calibrating structured light | |
| CN112465960A (en) | Dimension calibration device and method for three-dimensional model | |
| CN105844701A (en) | Sequential-image three-dimensional modeling method | |
| Yilmazturk et al. | Geometric evaluation of mobile-phone camera images for 3D information | |
| Aati et al. | Comparative study of photogrammetry software in industrial field | |
| CN112132900A (en) | Visual repositioning method and system | |
| Wenzel et al. | High-resolution surface reconstruction from imagery for close range cultural Heritage applications | |
| CN114792345A (en) | Calibration method based on monocular structured light system | |
| CN104614372B (en) | Detection method of solar silicon wafer | |
| Teppati Losè et al. | Preliminary evaluation of a commercial 360 multi-camera RIG for photogrammetric purposes | |
| Kampel et al. | Virtual reconstruction of broken and unbroken pottery | |
| Kampel et al. | Profile-based pottery reconstruction |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| C17 | Cessation of patent right | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20110504 Termination date: 20140428 |