JPH11259633A - Adjusting device of stereoscopic camera - Google Patents

Abstract

PROBLEM TO BE SOLVED: To easily adjust the optical position of the stereoscopic camera with high precision and low-cost constitution. SOLUTION: The camera stays of a main camera 2 and a subcamera 3 are fixed in the center to an adjustment rack, images of distant and near patterns are picked up and processed by a correction arithmetic device 50. Horizontal and vertical translation correction and rotation correction values based upon a rotation adjustment value for making the horizontal lines of the main camera 2 parallel to the reference line of the stereoscopic camera and an error in the correspondence position of the image of the subcamera 3 based upon the image of the main camera 2 are calculated. Then the mount of the main camera 2 is mechanically rotated on the optical axis at a rotary stage and an affine transforming circuit 36 geometrically transforms the image of the subcamera 3 to adjust the optical position of the stereoscopic camera.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a stereo camera adjusting device for adjusting an optical position of a stereo camera.

[0002]

2. Description of the Related Art In general, as a three-dimensional measurement technique using images, a correlation between a pair of images obtained by imaging an object from different positions with two cameras (stereo cameras) is obtained, and a stereo camera is attached based on parallax for the same object. 2. Description of the Related Art There is known image processing by a so-called stereo method in which a distance is obtained by the principle of triangulation using camera parameters such as a position and a focal length.

In the image processing by the stereo method, two image signals obtained from a stereo camera are sequentially shifted and superimposed to obtain a position where the two image signals coincide with each other. It is desirable that there is no deviation other than parallax, and it is important to adjust the optical position of the stereo camera.

For this reason, Japanese Patent Application Laid-Open No. 5-157557 discloses a holding member for connecting and holding a pair of video cameras.
Parallel adjustment means for adjusting the arrangement of the pixels of the image sensor of one video camera to be parallel to the arrangement of the pixels of the image sensor of the other video camera; the optical axis of one video camera and the optical axis of the other video camera There is disclosed a technology in which an optical axis adjustment member is provided for adjusting the camera so as to be parallel to each other, and the correlation between the two cameras is mechanically adjusted and maintained.

However, once a fixed stereo camera is displaced due to aging, conventionally,
It must be readjusted mechanically, which requires not only a complicated operation but also a long time for the readjustment, and mechanical adjustment has a limit in securing accuracy.

[0006] To address this, the applicant has first
Japanese Patent Application No. Hei 9-117268 proposes a technique for performing affine transformation on an image in accordance with a shift in the optical position of a stereo camera, thereby making electrical adjustment without mechanical adjustment. The optical position can be precisely adjusted to a level that is difficult to adjust mechanically, and a deviation due to aging after the adjustment can be easily readjusted.

[0007]

In the technique proposed by the present applicant, an image on the comparison side is corrected by an affine transformation circuit with respect to an image used as a reference in stereo processing. If the horizontal line of the camera on the side is inclined with respect to the base line, the image on the reference side must also be corrected by the affine transformation circuit, so the signal processing circuit of each camera has an affine transformation circuit. There is a need.

Therefore, there is a problem that the product cost is increased due to the complicated circuit configuration and the increase in the number of parts.
In the case where both images are corrected, the correction process takes a long time, which may hinder real-time image processing.

SUMMARY OF THE INVENTION The present invention has been made in view of the above circumstances, and has as its object to provide a stereo camera adjusting apparatus which can easily and accurately adjust an optical position with a low-cost configuration.

[0010]

According to the first aspect of the present invention,
A stereo camera adjusting device for adjusting an optical position of a stereo camera in which a first camera and a second camera are arranged on a camera stay at a set interval, and wherein the first camera has an optical axis. Means for rotating about the center and mechanically adjusting the optical position;
Means for geometrically converting an image by horizontal and vertical translational and rotational corrections based on the image of
A rotation adjustment value for the first camera is calculated so that a horizontal line of the camera is parallel to a base line of the stereo camera, and an image conversion value for the second camera is converted to a first image in an image of the second camera. Means for calculating based on the error of the corresponding position with respect to the image of the camera.

According to a second aspect of the present invention, in the first aspect of the present invention, the first camera picks up a pattern in which the horizontal translation correction amount with respect to the image of the second camera is arranged in a far and near direction. The calculation is performed based on the amount of positional deviation of each pattern between the image of the second camera and the image of the second camera, and the distance of each pattern.

According to a third aspect of the present invention, in the first aspect of the present invention, the rotation adjustment of the first camera is fitted to a fixing portion for fixing the camera stay and an outer periphery of a mount of the first camera. In addition, it is characterized in that the mounting is performed by an adjustment stand having a rotation stage for rotating the mount about the optical axis.

That is, according to the present invention, the first camera is rotated about the optical axis with a rotation adjustment value such that the horizontal line of the first camera is parallel to the base line of the stereo camera, and the image of the second camera is rotated. With the first in the image of the second camera.
The optical position of the stereo camera is adjusted by geometrically performing image conversion by horizontal and vertical translation correction and rotation correction calculated based on the error of the corresponding position with respect to the image.

In this case, as described in claim 2, the horizontal translation correction amount with respect to the image of the second camera is determined by comparing the image of the first camera with the image of the pattern arranged at a distance and the distance with the image of the second camera. And the distance between each pattern and the position of each pattern in the camera image.

The rotation of the first camera may be adjusted by fixing the camera stay to the adjustment stand and rotating the mount of the first camera about the optical axis by the rotation stage. Can be done by

[0016]

Embodiments of the present invention will be described below with reference to the drawings. 1 to 14 relate to an embodiment of the present invention, FIG. 1 is a configuration diagram of an image adjustment device, FIG. 2 is a block diagram of an affine transformation circuit, FIG. 3 is a perspective view of a stereo camera, and FIG. 5 is a perspective view of the rotation adjusting device from above, FIG. 6 is a perspective view of the rotation adjusting device from below, FIG. 7 is a cross-sectional side view of a main part of the rotation adjusting device, and FIG. FIG. 9 is a partial front view of the apparatus with an adjuster removed, FIG. 9 is an explanatory diagram showing the timing of image capture and inverse affine transformation, FIG. 10 is a flowchart of adjustment processing,
11 is an explanatory diagram showing pattern positions in a reference image and a comparative image, FIG. 12 is an explanatory diagram showing a rotation angle in affine transformation of a comparative image, FIG. 13 is an explanatory diagram showing a deviation of a horizontal line of a reference camera from a base line, FIG. 14 is an explanatory diagram of the camera system viewed from the side.

In FIG. 3, reference numeral 1 denotes a stereo camera comprising two cameras synchronized with each other and having a variable shutter speed. One camera 2 captures a reference image for stereo processing, As the other camera 3 as a sub-camera for capturing a comparative image at the time of stereo processing, both cameras 2 and 3 are arranged at a predetermined interval on a camera stay 4 formed by processing a thick plate. The camera stay 4 is provided with mounting holes 4b at five locations: a central portion between the cameras 2 and 3 and a diagonal line centered on the central portion.

The camera stay 4 has a mounting window 4 projecting from the front surface of the mount 5 of each of the cameras 2 and 3 and receiving a boss 5a (see FIG. 7) to which a lens barrel 9 is mounted.
a is drilled. As shown in FIG. 4, at four places around the mounting window 4a, long holes 4c are drilled at equal intervals, and a mount 5 is attached to the camera stay 4 via screws inserted into the long holes 4c. It is fixed. In addition, mount 5
The rotation of ± 6 ° about the optical axis is allowed by the long hole 4c.

Further, a circuit board 7 (see FIG. 7) for mounting an image pickup device 6 such as a CCD at the focal position of a lens optical system is provided on the rear surface of the mount 5 of each of the cameras 2 and 3 with a knock pin (see FIG. 7) at least on a diagonal line. (Not shown), and a circuit board 8 in which peripheral circuits such as a CCD drive circuit are incorporated is fixed via a spacer 8a.

In the above stereo camera 1, the optical position is adjusted by mechanically rotating one axis about the optical axis of one camera and correcting the electric image of the other camera. The optical position can be easily and accurately adjusted with a simple circuit configuration, and the total cost can be reduced.

In this case, it is desirable to mechanically adjust the main camera 2 serving as a reference for stereo processing and to electrically correct the image of the sub-camera 3. In the present embodiment, the rotation adjustment device 11 shown in FIG. The rotation of the main camera 2 around the optical axis is adjusted so that the horizontal line of the main camera 2 is parallel to the base lines of the cameras 2 and 3, and the image of the sub camera 3 is shown in FIG. It is electrically corrected by the image adjustment device 20.

First, the rotation adjusting device 11 will be described with reference to FIGS.
This will be described with reference to FIG. An adjustment frame 12a is erected on the adjustment base 12 of the rotation adjustment device 11, and an adjustment surface 12b formed on a side surface of the adjustment frame 12a is provided.
A bracket 12c, which is an example of a fixing portion, protrudes from the center of the bracket. In the bracket 12c, of the mounting holes 4b drilled in the camera stay 4, the central portion and the mounting hole 4b drilled at the lower right of FIG. 4 are screwed.

Further, a rotary stage 13 is mounted on the adjustment surface 12b on the side opposite to the main camera 2. As shown in FIGS. 7 and 8, a pin 15 is screwed into the rotation center of the rotary stage 13, and the rotation center of the adjuster 14 is engaged with the pin 15 and positioned. The adjuster 14 is formed in a cylindrical shape with an open front, and the back of the adjuster 14 is
It is screwed to. The rotary stage 13 is rotated by a micrometer 16 provided on one side thereof. In the present embodiment, a rotary stage having a resolution of 0.1 ° or less is employed.

The outer periphery of the mount 5 of the main camera 2 is fitted to the inner periphery of the open end of the adjuster 14.
The inner circumference of the opening of the adjuster 14 and the outer circumference of the mount 5
The adjuster 14 is formed with a high-precision hammer-eye size, and the optical axis of the main camera 2 and the rotational center of the adjuster 14 are accurately matched.

A pair of rotation adjusting pins 17 are screwed on the diagonal line of the opening end of the adjuster 14, and the tips of the rotation adjusting pins 17 press the outer periphery of the mount 5, and Are fixed to the adjuster 14. A slit 14a is formed in the outer periphery of the adjuster 14 at a position shifted in phase from the rotation adjusting pin 17 along the axial direction from the opening end side. This slit 14a
This is for taking out a cable (not shown) extending from the circuit board 8 to the outside.

On the other hand, the image adjusting device 20 for performing image correction on the comparison image of the sub camera 3 has a function as shown in FIG.
An image input device 30 for processing images picked up by both cameras 2 and 3 and outputting the processed image to a stereo processing device (not shown) at the subsequent stage, and a correction value for a comparison image which is connected to the image input device 30 during adjustment. And a correction operation device 50 for performing the operation.

The image input device 30 includes cameras 2, 3
Analog interfaces 31 and 3 for adjusting the analog image to the input range at the subsequent stage corresponding to each system
2. A / D converters 33 and 34 for converting an analog image into a digital image having a predetermined luminance gradation (for example, 256 gray scales) are provided. An input image memory 35 for temporarily storing an image, and an affine transformation circuit 36 for performing geometric transformation such as image rotation and translation for the comparison image stored in the input image memory 35 are provided.

The affine transformation circuit 36 has an internal configuration as shown in FIG. 2, and has an image memory data interface 37 for writing data to and reading data from the input image memory 35, and an address of the input image memory 35. An image memory address interface 38 for specifying, an input image write address generation circuit 39 for generating an address for writing image data from the A / D converter 34 to the input image memory 35, and the input image memory 3
5, an inverse affine transformation read address generation circuit 40 for generating an address at the time of performing geometric transformation of an image by performing inverse affine transformation, performs linear interpolation on the data read by inverse affine transformation And an interpolation operation circuit 41 for outputting converted image data.

That is, in the geometric transformation of the image by the affine transformation circuit 36, the original image before the transformation and the transformed image are both digital images in which pixels are arranged on a square lattice. The density value of a pixel is given by a density value at a corresponding pixel position obtained by performing inverse affine transformation on an original image. In general, the corresponding pixel position of the original image by the inverse affine transformation is not an integer pixel position, and no corresponding pixel exists in the original image.
Linear interpolation is performed using the density values of the pixels, and the density values of the pixels are obtained without gaps on the converted image.

As shown in FIG. 9, the affine transformation circuit 36
Then, each of the cameras 2 and 3 during the sampling period of the field signal
, For example, an image signal such as an NTSC video signal output from each of the cameras 2 and 3 in synchronization with the horizontal synchronizing signal and the vertical synchronizing signal. Convert.

That is, the analog interface 3
The gains and offsets of the image signals from the cameras 2 and 3 are adjusted in accordance with the input ranges of the A / D converters 33 and 34 by the A / D converters 1 and 32, and the A / D converter 34 on the sub camera 3 side performs A / D conversion. The obtained digital image data is stored in the input image memory 35 in accordance with the address generated by the input image write address generation circuit 39 of the affine conversion circuit 36, and the inverse affine conversion of the affine conversion circuit 36 is performed from the input image memory 35 in the next field section. The density data of the address generated by the conversion read address generation circuit 40 is read. Then, an interpolation operation is performed on the density data by the interpolation operation circuit 41 of the affine conversion circuit 36 to output a converted image.

The correction operation device 50 includes a CPU 51,
A computer including image memories 52 and 53, a keyboard 54, a CRT display 55, and the like. The computer is connected to the A / D converters 33 and 34 of the image input device 30 via the image memories 52 and 53, and adjusts the rotation adjustment value of the main camera 2. Is displayed on the CRT display 55 and the rotation adjusting device 11 is displayed.
Instruct the operator to make mechanical adjustments, and output the image correction value of the sub camera 3 to the affine transformation circuit 36 as setting data.

Next, the procedure for adjusting the stereo camera 1 will be described. The adjustment of the stereo camera 1 is performed by setting the stereo camera 1 on the rotation adjustment device 11 and connecting the image input device 30 and the correction operation device 50.

The stereo camera 1 is set on the rotation adjusting device 11 through the mounting holes 4b formed in the center and lower right of the mounting holes 4b formed in the center of the camera stay 4. Then, the camera stay 4 is rotated by the rotation adjusting device 11.
Is fixed to the front surface of a bracket 12c protruding from the side surface of the adjustment gantry 12.

In this case, since the camera stay 4 has only a central portion supported by the bracket 12c, no stress is applied to the cameras 2 and 3 disposed on the left and right sides, and the camera stay 4 is not deformed. Is prevented. Further, since the camera stay 4 is merely fixed with two screws, the detachment is easy.

The outer periphery of the mount 5 on the side of the main camera 2 fixed to the camera stay 4 is housed in the opening end of the adjuster 14 provided in the rotation adjusting device 11, and the rotation adjusting pin 17 is used. Fix it. Since the inner periphery of the open end of the adjuster 14 and the outer periphery of the mount 5 are formed with high-precision hammer-eye dimensions, the optical axis of the main camera 2 and the rotation center of the adjuster 14 must be Are matched exactly.

Next, the circuit board 8 of the main camera 2
Is pulled out through a slit 14 a formed on the outer periphery of the adjuster 14 and connected to the image input device 30. In addition, the mount 5 of the main camera 2 is set in a temporary fixing state in which the rotation direction is allowed by a screw in the long hole 4 c formed in the camera stay 4.

Then, in front of the stereo camera 1, a target having an appropriate adjustment pattern is set at one location in the vicinity and two targets in the distance.
The stereo camera 1 is installed at various locations and images of these patterns are taken. The distant target is installed at a position about 10 times apart from the near target, for example, and the distance from the stereo camera 1 is accurately measured. When the above preparations are completed, the correction calculation device 50 connected to the image input device 30 starts the adjustment processing program shown in FIG. 10 and processes the image of the pattern captured by the stereo camera 1 to calculate the adjustment value. .

In this adjustment processing program, the steps
In S100, a display for instructing input of the position of each target in the reference image is performed on the CRT display 55, and a loop waiting for input is performed in step S110. When the operator measures the position of the target in this reference image and inputs the position of each target from the keyboard 54 or an input device such as a mouse (not shown), the process proceeds from the input waiting loop to step S120, where the target in the reference image is templated. Then, a process of obtaining the position of each target in the comparison image by well-known template matching is performed.

As a result, as shown in FIG. 11, in a coordinate system in which the origin is at the upper left of the image, the horizontal direction is X, and the vertical direction is Y coordinate, two targets # 1R and # 2 at the far side of the reference image.
For the position coordinates (Xr1, Yr1), (Xr2, Yr2) of R and the position coordinates (Xr3, Yr3) of one target # 3R near, two targets # 1C, # at a distance in the comparison image. 2C position coordinates (Xc
(1, Yc1), (Xc2, Yc2) and the position coordinates (Xc3, Yc3) of one nearby target # 3C are obtained.

In the following step S130, the correspondence between the position coordinates (Xr1, Yr1) of one target on the reference image side, for example, target # 1R on the far left side in FIG. Coordinates of the target # 1C on the far left (Xc1,
The difference (Yr1−Yc1) of the Y component from Yc1) is obtained, and the difference (Yr1−Yc1) of the Y component is used as the translation correction amount in the Y direction, ie, the image translation conversion value in the Y direction. In addition, the image (comparative image) of the sub camera 3 is translated in the Y direction.

Next, the process proceeds to step S140, where X of the comparison image
The direction translation correction amount is obtained, and this X direction translation correction amount is added to the existing value of the affine transformation circuit 36 to convert the image of the sub camera 3 into the X direction.
Translate in the direction. The X-direction translation correction amount can be given by a deviation amount Z0 between the reference image and the comparative image at infinity, and the distance d1 to the distant target, the distance d3 to the near target, the distant reference image and the comparative image And the displacement Z0 at infinity is expressed by the following equation (1) using the displacement Z1 between the reference image and the comparative image, and -Z
Let 0 be the amount of translation of the comparative image in the horizontal direction (X direction).

Z 0 = (d 3 · Z 3 -d 1 · Z 1) / (d 1 -d 3) (1) Originally, the matching points in the reference image and the comparison image differ only in the horizontal X coordinate at which parallax is detected. Y coordinates must be the same. Therefore, in step S130, the Y coordinate position of the reference image and the comparison image is matched with one of the two distant targets, and in step S140, the X coordinate position of the comparison image is changed to the original X coordinate position. It is adjusted to the position where parallax is to be detected.

In the next step S150, the comparative image is rotated to make the Y coordinate position of the other target coincide between the reference image and the comparative image. That is, as shown in FIG. 12, with the position coordinates (Xr1, Yr1) of the target # 1R on the far left side of the reference image as the rotation center, the Y coordinate value Yc2 of the target # 2C on the far right side of the comparison image and the far end of the reference image are displayed. Y coordinate value Y of right target # 2R
Rotate the comparison image so that r2 matches. here,
Coordinates (Xc2, Yc) at a rotation angle θ around the coordinates (Xr1, Yr1)
When (2) is rotated, the coordinates (Xc2 ′, Yc2 ′) after the rotation are represented by the following equation (2).

[0045] The above equation (2) is based on the assumption that the Y coordinate value Yc2 ′ of the target # 2C in the rotated comparison image matches the Y coordinate value Yr2 of the target # 2R in the reference image, and the Y component is set as Yc2 ′ = Yr2. When only these are arranged, they are expressed by the following equation (3).

Yr 2 −Yr 1 = (Xc 2 −Xr 1) · sin θ + (Yc 2 −Yr 1) · cos θ (3) Further, in the above formula (3), Yr 2 −Yr 1 = A, Xc 2 −
Xr1 = B, Yc2-Yr1 = C, cos θ = ±
By substituting (1−sin ² θ) ^1/2 and solving for sin θ, the following equation (4) is obtained.

Sin θ = (A · B ± C · (B ² + C ² −A ² ) ^1/2 ) / (B ² + C ² ) (4) The value of the above equation (4) is Yr 2 = Yc 2 and A = C, that is, when the Y coordinate of the target # 2R on the far right side of the reference image already coincides with the Y coordinate value of the target # 2C on the far right side of the comparison image, there is no need for rotation. From the following equation (5), it can be seen that the sign of the second term of the numerator of the above equation (4) must be negative.

Sin θ = B · (A ± C) / (B ² + C ² ) = 0 (5) Therefore, the rotation angle θ can be finally obtained by the following equation (6). Is added to the existing value of the affine transformation circuit 36 as an image rotation conversion value about the coordinates (Xr1, Yr1), and the image of the sub camera 3 (comparison image) is rotated.

Θ = sin ⁻¹ (A · B−C · (B ² + C ² −A ² ) ^1/2 ) / (B ² + C ² ) (6) By rotating the comparative image, the vicinity of the comparative image is obtained. Target #
3C also rotates, but as shown in FIG.
3 is not parallel to the horizontal line of the main camera 2 as the reference camera, and the horizontal line of the main camera 2 is inclined by an angle φ with respect to the base line, the near target of the reference image FIG. 1 shows the relationship between the Y coordinate value Yr3 of # 3R and the Y coordinate position Yc3 of the near target # 3C of the comparative image after rotation.
A difference ΔYm−ΔYs as shown in FIG.

Therefore, in the next step S160, the difference ΔYm
In order to eliminate -ΔYs, a rotation angle φ about the coordinates (Xr1, Yr1) of the reference image is calculated. This rotation angle φ is as shown in FIG.
As can be seen from FIG. 14, it can be obtained from the base line length B of the cameras 2 and 3 and the deviation ΔY between the center of the reference image and the center of the comparison image on the focal plane.
From the geometrical relationship when the camera system is viewed from the side as shown in the figure, it can be obtained using the focal length f, the distance d1 to the distant target, the distance d3 to the near target, and the difference ΔYm−ΔYs. it can.

That is, using the difference ΔYm−ΔYs between the deviation ΔYm from the center of the image in the reference image of the near target formed on the imaging surface and the deviation ΔYs from the center of the image in the comparison image, the following equation (7) is used. ΔY can be obtained, and from this deviation ΔY and the base line length B, the rotation angle φ can be finally obtained by the following equation (8).

ΔY = (ΔYm−ΔYs) · d1 · d3 / (f · (d1−d3)) (7) φ = tan ⁻¹ (ΔY / B) = tan ⁻¹ ((ΔYm−ΔYs) · d1 D3) / (Bf (d1-d3)) (8) Next, the process proceeds to step S170 to check whether or not the rotation angles θ and φ are sufficiently small in accuracy and within the allowable value. If it does not reach the allowable value, the process proceeds to step S180, in which the rotation angle φ of the main camera 2 is displayed on the CRT display 55, and rotation adjustment is instructed.

In accordance with the display on the CRT display 55, the operator operates the micrometer 16 provided on the rotation adjusting device 11 to rotate the rotary stage 13, and
The mount 5 of the main camera 2 fixed to the rotary stage 13 via the adjuster 14 is rotated.

In this case, since the rotating stage 13 has a resolution of 0.1 °, the rotation can be adjusted with necessary and sufficient accuracy, and highly accurate adjustment can be efficiently performed. Furthermore, since the center of rotation of the rotary stage 13 during rotation adjustment and the center of the adjuster 14 are aligned via the pin 15 protruding from the center of the adjuster 14, the optical axis does not shift, and , Translational movement of the camera can be prevented.

Then, the main camera 2 is rotated around the optical axis by the rotation angle φ calculated in step S160, and an input to the effect that the adjustment has been performed is input from the keyboard 54 or an input device such as a mouse (not shown). The process returns from step S190 to step 120 to calculate the position of the corresponding target in the comparison image and perform the same processing (and operation).

When the above adjustments are repeated and the rotation angles θ and φ become sufficiently smaller than the allowable values, the translation correction amount in the Y direction,
The X-direction translation correction amount and the rotation angles θ and φ are fixed as the final image conversion values of the affine conversion circuit 36, and the effect of the completion of the adjustment is displayed on the CRT display 55, and the process ends. Thus, the worker completes the adjustment of the stereo camera 1 by fastening and fixing the screws temporarily fixing the mount 5 of the main camera 2 to the camera stay 4.

[0057]

As described above, according to the present invention, the first camera is mechanically rotated about one optical axis around the optical axis.
Since the image of the second camera is electrically corrected, the optical position of the stereo camera can be easily and accurately adjusted with a simple configuration, and the configuration of the image input circuit can be simplified. In addition, excellent effects can be obtained such as reduction in parts cost.

[Brief description of the drawings]

FIG. 1 is a configuration diagram of an image adjustment device.

FIG. 2 is a block diagram of an affine transformation circuit;

FIG. 3 is a perspective view of a stereo camera.

FIG. 4 is a front view of the camera stay.

FIG. 5 is a perspective view of the rotation adjusting device as viewed from above.

FIG. 6 is a perspective view of the rotation adjusting device as viewed from below.

FIG. 7 is a sectional side view of a main part of the rotation adjusting device.

FIG. 8 is a partial front view of the rotation adjusting device with an adjuster removed.

FIG. 9 is an explanatory diagram showing timings of image capture and inverse affine transformation.

FIG. 10 is a flowchart of an adjustment process.

FIG. 11 is an explanatory diagram showing the positions of targets in a reference image and a comparison image.

FIG. 12 is an explanatory diagram showing a rotation angle in affine transformation of a comparative image.

FIG. 13 is an explanatory diagram showing a deviation of a horizontal line of a reference camera from a base line.

FIG. 14 is an explanatory diagram of the camera system viewed from the side.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 ... Stereo camera 4 ... Camera stay 11 ... Rotation adjustment device 12 ... Adjustment stand 12c ... Fixed part (bracket) 13 ... Rotation stage 20 ... Image adjustment device 30 ... Image input device 36 ... Affine conversion circuit 50 ... Correction arithmetic device

──────────────────────────────────────────────────の Continued on the front page (51) Int.Cl. ⁶ Identification symbol FI // G01B 11/24 101 G06F 15/62 415

Claims (3)

[Claims] 1. A stereo camera adjusting device for adjusting an optical position of a stereo camera, wherein a first camera and a second camera are arranged on a camera stay at a set interval, and Means for rotating the camera about the optical axis to mechanically adjust the optical position; and correcting horizontal and vertical translation and rotation correction of the image of the second camera with respect to the first image. Means for geometrically converting the image by means of: calculating a rotation adjustment value for the first camera so that a horizontal line of the first camera is parallel to a base line of the stereo camera; Means for calculating an image conversion value based on an error of a corresponding position of the image of the second camera with respect to the image of the first camera. 2. The horizontal translation correction amount for the image of the second camera is calculated by calculating the amount of translation correction of each pattern in a first camera image and a second camera image obtained by capturing a pattern arranged in the far and near directions. 2. The stereo camera adjustment device according to claim 1, wherein the calculation is performed based on a positional shift amount and a distance between each pattern. 3. The rotation adjustment of the first camera is performed by a fixing portion that fixes the camera stay, and a rotation that fits around an outer periphery of a mount of the first camera and rotates the mount about an optical axis. 2. The stereo camera adjustment device according to claim 1, wherein the adjustment is performed by an adjustment stand having a stage.