TITLE LENS ALIGNMENT SYSTEM FOR SOLID STATE IMAGER
Field of Invention The invention relates to a method and apparatus for aligning a lens relative to an imager so that the optical center of the lens is aligned with the optical center of the imager.
Background of the Invention
There are several situations in which it is important that the optical center of a lens be in exact optical alignment with the optical center of the imager. One such application involves stereo processing of images from two imagers spaced a known distance apart. By using various processing techniques these images can be used to determine the three dimensional position of an object or even a person's right or left eye. Typically, a charge couple device (CCD) or a complementary metal oxide semiconductor (CMOS) is used as the imager.
The mounting of CCD (or CMOS) sensor chips to their package is inherently inconsistent due to manufacturing and assembly (chip placement) tolerances. This is especially apparent over different manufacturing lots where one can observe variations in the rotation and the tilt of the chip with respect to its bonding pad within its package. Additionally, the mounting placement of the package to a printed circuit board may not be very accurate unless costly procedures are used. Thus, there is difficulty in using reference holes or fiducial marks on a camera circuit board to align the optical center of a lens with that of the CCD chip. Yet, there is a critical need for alignment of the center of the
camera sensor to its lens for proper operation of at least some stereo range finding algorithms.
It is much easier to create an integral camera/lens package in which the optical centers of the camera and the lens are aligned than to try to mount the camera and lens separately in a device. There is a need for a method and device to perform the alignment of the camera board to its lens in an offline process and then drop this assembly into another device so that the camera axis and lens axis are properly aligned with mechanical structures of the device such as a pair of stereo camera mounts. This approach eliminates the need to align each camera and lens pair separately, making a more consistent assembly and also allows interchange of camera modules. The consistent alignment of lens camera assemblies to their mounts facilitates software controlled calibration of stereo vision systems, since the position and angular offset between camera centerlines can be maintained within an acceptable tolerance thereby reducing the cost usually associated with the manual alignment and calibration of stereo camera pairs. Use of a lens camera package also permits easier replacement of the imager in a system should there be a need to do so.
Summary of Invention We provide a device and technique for aligning the center of a CCD sensor to the optical centerline of its respective lens to create a camera lens package that can be used in a more complicated electro-optical system. The approach takes into account the mounting tolerance issues of the die to package and package to circuit board. The final result is a
camera and lens whose centers are accurately referenced to a mounting structure. The camera board contains the imaging elements but is not used for any mounting reference.
To align an imager relative to a lens so that an optical center of the imager is aligned with an optical center of the lens we first place the lens in a lens support so that the optical center of the lens is at a known point relative to the lens support. Then the lens support structure is placed in a secondary structure in front of a target device having a pair of crosshairs. The center of the secondary structure has been mechanically aligned with the center of the crosshairs and the centerline of the lens in the lens support is coincident with the centerline of the secondary structure. The target device has at least two alignment holes which are aligned with alignment rods that extend from the secondary structure holding the lens support. These alignment rods ensure that the target and secondary structure remain in alignment and are parallel to the optical axis of the lens. We prefer to provide a plate to which the lens support is removably attached. The target is placed on rails to allow it to be slid toward and away from the lens support during the alignment procedure with the alignment rods fixing the relationship between the target and secondary structure.
At this point there will be a line perpendicular to the front face of the lens support which passes through the optical center of the lens and a center of the pair of crosshairs. Then the imager is positioned opposite the back face of the lens support and an electronic crosshair is generated at the optical center of the imager array. The imager is then moved in the x, y, and theta directions to a position where the generated pair of crosshairs of the image overlays the pair of crosshairs of the target when viewed on a monitor connected to
the imager. At that point the imager and the lens are aligned. To complete the procedure the imager is affixed to the lens support to form the assembled camera lens package. The aligned camera lens package is now ready for installation in another device.
Other objects and advantages will become apparent from a description of certain present preferred embodiments illustrated in the drawings.
Brief Description of the Figures
Figure 1 is a perspective view of a camera lens package which can be built using the present alignment method and apparatus.
Figure 2 is an exploded view of the camera lens package shown in Figure 1.
Figure 3 is a side view of a present preferred alignment fixture.
Figure 4 is a front view of a present preferred target.
Figure 5 is a top plan view of the camera lens package with the attachment portion of the preferred alignment fixture.
Figure 6 is a side view of the camera lens package and alignment fixture shown in Figure 5.
Figure 7 is a front view of the alignment fixture shown in Figures 5 and 6.
Figure 8 is a perspective view of the alignment fixture shown in Figures 5, 6 and 7.
Description of the Preferred Embodiments The object of the present invention is to create a camera lens package 1 such as is shown in Figures 1 and 2. This package consists of a lens 4 within a connector 20 to
which an imager 6 is attached. The imager 6 is comprised of a CCD array package 28 with a filter 24 held within a frame 23 and attached to a housing 26 and camera board 19. Typically, integrated circuit packs 13 which supply control and timing for the CCD array package 28 will be attached to the camera board 19. In a present preferred package, the lens has a transparent portion 14 contained in a housing 15 which has a threaded elongated stem 16 that is screwed into mechanical connector 20. Connector 20 is frusto-conical and fits within a tapered hole 9 in a rigid support 2 which we also call a lens support. Hole 9 is positioned so that its center has coordinates which are known relative to the base and edges of the upright. A key (not shown) extends into tapered hole 9 and mates with slot 21 on the outer surface of the connector 20. The smaller diameter end of the connector 20 has external threads 22. Conical connector 20 is inserted into tapered hole 9 of the rigid support 2 from the opposite face. A nut 17 is tightened around external threads 22 of the connector 20 to retain the connector 20 within tapered hole 9. The key and slot prevent rotation of the mechanical connector 20 as the nut is tightened or as the lens housing 15 is screwed in and out of the mechanical connector to facilitate focus adjustment. A CCD array package 28 with integral connectors 29 is mounted on the camera circuit board. The array package may be a surface mount device or may be mounted in a IC socket on the circuit board. A pair of screws 18 holds housing 26, filter 24 and filter frame 23 over the CCD array package 28. When the imager is properly aligned with the lens, a second pair of screws 37 with washers 38 will extend through the oversized holes in camera board 19 and oversized holes 25 in housing 26 into tapped holes (not shown) in the rear of connector 20. In this way, the imager 6 is attached to the connector 20 carrying lens 4. We prefer to
also provide an adhesive on rear surface of the connector 20 to more securely hold the imager on the connector 20.
The imager must be mounted so that the optical axis of the lens and the optical axis of the CCD imager are colinear. To perform this alignment, a rigid mount or lens support 2 carrying the lens 4 and connector 20 are mounted onto a plate 12. The connector 20 is secured to the support 2 with a nut 17. Then a target 40 is provided a selected distance in front of the lens 4 using an alignment fixture 30 shown in Figure 3. The alignment fixture 30 has a mounting plate or base 31 on which the support plate 12 and attached lens support 2 are placed. A set of guides 34 define the location at which the support plate 12 is placed on the base 31. If desired a slot could be cut in the base 31 with the sides of the slots acting as the guides. Alternatively a single guide could be used provided it is dimensionally stable. When the plate 12 and lens support 2 are placed between the guides 34 the rear face of the lens support 2 will be perpendicular to a set of rails 35 on the opposite end of the base 31 which carry the target 40.
Two removable alignment rods 32 with alignment pins 33 are placed perpendicular to the front face of the lens support 2. These rods will also be perpendicular to a horizontal diameter at through the hole 9 in the lens support 2. The target 40 is carried on carts 42 which have rollers 44 that ride on rails 35. Each cart has a hole 45 to interface to the alignment pins 33. The target has an xy pair of crosshairs 46 on its face. The xy pair of crosshairs 46 is located on the target 40 such that if the target is adjusted so that the two alignment holes 45 interface the alignment pins 33, the xy pair of crosshairs is on the centerline of the tapered hole 9 of the rigid mount 2. The target 40 contains adjustments
47 and 48 for x, y positioning of the two alignment holes (and hence the xy pair of crosshairs to the centerline of the tapered hole 9). These adjustments move the horizontal line of the cross-hairs up and down relative to the carts and left and right relative to the carts. The crosshairs 46 on target 40 are held parallel to the front of the lens and lens support 2 by the rails 35. Sliding clamps or sliders 49 are provided which slide along the rails 35 and then are tightened to hold the target at a desired position on the rails. Then the pair of crosshairs 46 is aligned in x and y (with the use of the mechanical adjustments 47 and 48) until the alignment pins 33 mate with the alignment holes 45. This process guarantees that the center of the crosshairs is on the center line of the lens support 2.
The alignment of the crosshair, 46 to the center of hole 9 can be accomplished either mechanically or optically. For the mechanical case, a device similar to mechanical connector 20 is fabricated but with a stiff rod approximately 20 inches long emanating from its center. Consequently, the rod would be colinear with the optical center line 8 of the lens shown in dotted line in Figure 3. The target 40 is moved into position with the two alignment rods mating with the lens support 2. The mechanical adjustors are then moved until the center of the crosshair, 46 is coincident with the tip to the stiff rod. By spring loading the rod along its centerline, some compliance is obtained and the length of the rod and alignment rods is not as critical. For optical alignment, a small laser diode is placed in the center of a device similar to mechanical connector 20. Once again target, 40 is moved into position and the two alignment rods are mated with the lens support 2. Next the laser diode is activated and the mechanical adjustors are used to move the crosshair
center so that it is coincident with the small dot produced by the laser from a beam which travels along line 8 in Figure 3.
Once the crosshairs are aligned to the center of hole 9, the target 40 position is locked in x and y, the alignment rods 32 are removed. Then the target is locked at a position approximately 20 inches in front of the lens support 2. Typically this distance will be from 15 to 25 inches. This alignment need only be done one time.
A second alignment fixture 50 shown in Figures 5 through 8 is used for aligning the CCD to the lens. This alignment device 50 will be placed on plate 12 in the location indicated by the box 50' shown in chain line in Figure 3 and is discussed in subsequent steps.
To align the CCD sensor with the lens, the camera board 19 with CCD array package 28 is attached to an alignment device 50 shown in Figures 5,6, 7 and 8. The alignment device 50 is attached to plate 12 and is able to move the camera 6 in the X, Y planes and can rotate the camera about the Z-axis. The Z-axis is in the direction of the optical axis of the lens. The attachment of the camera 6 to the alignment device 50 is through a mounting hole pattern in the camera board 19. The alignment device 50 is used to position the camera board 19 near the back focal point of the lens 4. At this point, the CCD panel is positioned at the nominal focal point of the lens 4 at the end of the mounting structure 52 of the alignment device 50 by screws 53 which mate with posts 51. In one embodiment the posts 51 are tapered and the circuit board is pushed onto the posts. The objective of the next steps is to align the center of the CCD array 28 to the crosshairs approximately 20 inches in front of the lens mounting structure. This is done whenever the camera package
will be used in a device in which the center of the CCD array must be aligned with the center of the lens. Based on manufacturing tolerance data from the CCD sensor manufacturer and the optical design requirements for such systems in which the camera package will be used, the adjustments made in this step should be the only adjustments required for alignment.
Electronics and a monitor are connected to the camera package through the camera's electronic interface connector 60 so that the target 40 containing the crosshairs may be imaged. If necessary, a focus adjustment is made by screwing the lens 4 in or out of the connector 20. Note that the CCD location is fixed in the Z direction with respect to the mounting structure and subsequent adjustments of the CCD location will not affect focus.
A pair of crosshairs are electronically generated. The crosshairs are synchronized to the actual pixel array of the CCD sensor. The crosshairs are set to correspond to the location of the horizontal and vertical center of the array. The electronic crosshairs will be displayed superimposed on the image generated by the CCD. At this point, the monitor should show both crosshairs, the electronically generated crosshairs corresponding to the center of the CCD array and the target crosshairs 46 corresponding (as best as possible) to the optical axis of the lens.
The goal of the alignment process is to have the crosshairs aligned to show what appears to be only one pair of crosshairs. When this occurs the optical axis of the lens is aligned with the center of the CCD sensor. The two crosshairs will not be aligned when the camera is first turned on. Two distinct pairs of crosshairs will be seen. The target
crosshairs have previously been positioned with one leg oriented vertically and one horizontally with respect to the final mounting orientation of the camera/lens assembly. The generated CCD crosshairs will likely be rotated and offset slightly with respect to the fixed target crosshairs.
The alignment device 50 to which the camera board 19 is attached uses three fine pitch screws 54, 55 and 56 for moving the camera board 19 carrying the CCD sensor into position. These screws 54, 55 and 56 permit movement of camera board 19. In the X direction, the Y direction and rotation about the Z-axis, respectively. The interface between the alignment device 50 and the camera board 19 is a flat spring 58, which picks up three or more fixturing holes and permits transmission of the screw movement to the camera board. The flat spring 58 also maintains contact between housing 26 and connector 20.
The following steps are followed for alignment. First, focus the crosshairs image on the CCD sensor. Then adjust rotation about the Z-axis making the crosshairs parallel. Next adjust sensor position in the horizontal axis until there is one vertical line and two horizontals. Then adjust sensor position in the vertical axis until there appears to be only one pair of crosshairs. The camera 6 can be moved relative to connector 20 because holes 25 in housing 26 and corresponding holes in circuit board 19 are larger in diameter than the diameter of screws 37. Finally, fix the focus and sensor mounting.
The lens position defining focus is fixed by applying a drop of glue or LOCKTITE adhesive or other adhesive on threads 16 where they meet the internal threads on connector 20. The CCD's x,y and rotary alignment may be fixed by tightening two set screws 37 or
by an adhesive. After the camera is attached to the mechanical connector 20 via screws 37 or adhesive or both the alignment is complete. The alignment device 50 is retracted from the camera board 19, nut 17 is removed, and the camera/lens assembly 1 is then removed from support 2. The assembly is now ready to mount into the final structure of the product.
If desired, the lens may remain in the lens support and the lens camera package could be placed in the product while in the lens support 2. If the lens support 2 is to become part of the product, the lens and camera could both be attached to the support 2, but not directly connected to each other. The mounting in the product duplicates the mounting in the alignment fixture and as stated previously does not rely on the camera circuit board for alignment or mounting.
The screw alignments in the alignment device 50 are currently manually performed. However, the adjustments may be motorized and with use of machine vision software may be completely automated. Hence, a vision system could detect the rotational and positional offset of the physical crosshairs, 46 with respect to the crosshairs corresponding to the center of the CCD array package. An error signal for x, y and rotation could be generated from the vision system and a controller could generate appropriate signals to the motors to drive the adjustments in the direction to minimize or zero the error signal. Either stepper or servo motors capable of moving in single step increments could be used to move adjustment screws 54, 55 and 56.
In the preferred embodiment a single lens was attached to the camera. However, those skilled in the art will recognize that a multiple lens assembly could be used in place
of the single lens. Furthermore, one or more filters could be attached to the lens or in the multiple lens assembly.
Although we have disclosed certain present preferred embodiments of our alignment method and apparatus it should be distinctly understood that our invention is not limited thereto but may be variously embodied within the scope of the following claims.