CN110166766B - Multi-line array CCD camera coplanar collinear imaging combined debugging method - Google Patents

Abstract

The invention discloses a co-planar and co-linear imaging combined debugging method for a multi-line array CCD camera, which is used for accurately adjusting the accurate position, the height, the front and back positions, the parallelism and the pitching angle of the camera one by one. The method solves the problem that the traditional co-planar and co-linear shooting combined debugging method of the multi-linear array CCD camera is difficult to meet the requirements of calibration efficiency and precision, and realizes high calibration efficiency and high precision.

Description

A joint debugging method for coplanar and collinear imaging of multi-line CCD cameras

technical field

The invention relates to the technical field of machine vision, in particular to a coplanar and collinear imaging joint debugging method of a multi-line array CCD camera.

Background technique

With the rapid development of the machine vision industry, when collecting images of large objects, it is necessary to use multiple linear CCD cameras to collect images at the same time. Therefore, the correct stitching of images collected by multiple linear CCD cameras plays an important role. A frame of image captured by a line-array CCD camera is a row of pixels. When using multiple line array CCDs, it is necessary to adjust the height, position and angle of the multiple line array CCD cameras, so that the images collected by the multiple cameras at the same time correspond to the same line on the object. Therefore, many experts propose a variety of methods to solve the joint debugging of multiple linear CCD cameras. The current line scan camera pose calibration methods are mainly divided into methods based on optical calibration and methods based on calibrators and image processing. The optical calibration method divides the measured object into several areas, each area is equipped with a set of independently adjustable line scan cameras and optical systems, and each camera is adjusted one by one to meet the measurement requirements. This type of method is mainly aimed at optical measurement requirements and has high accuracy, but has a single adjustment function, relies too much on the performance of the optical system and hardware accuracy, and is costly. The main research idea of the method based on calibrator and image processing is to design a specific calibration plate and pattern, and adjust it with the combination of image processing algorithm and motion control mechanism. Such methods have certain flexibility, but have limitations such as high motion control accuracy, complex algorithms for centroid extraction and fitting, and the need for accurate dynamic imaging, and are not reliable enough for industrial applications. Aiming at the existing methods that rely too much on hardware and complex image processing algorithms, the present invention proposes a simple and efficient calibration method in combination with practical industrial applications, taking into account the requirements of calibration efficiency and accuracy.

SUMMARY OF THE INVENTION

Based on the technical problems existing in the background art, the present invention proposes a coplanar and collinear imaging joint debugging method for a multi-line CCD camera.

The technical scheme adopted in the present invention is:

A coplanar and collinear imaging joint debugging method for a multi-line array CCD camera, characterized in that it includes the following steps:

(1) Camera height and camera position

(1) Establish the imaging system spatial coordinate system to describe the spatial position information of the imaging system. Let dx, Ly, and Lz be the three-dimensional reference lines of the imaging system spatial coordinate system, respectively; Lx is the left edge of the detection object TFT-LCD glass substrate , Ly is the horizontal reference line drawn when the light source is adjusted, Lz is a vertical reference line that passes through the intersection of the straight lines Lx and Ly and is perpendicular to the reference plane where the detection object is located;

(2) When the standard magnification of the lens is n, the working distance of the lens is h mm, that is, the distance from the detection reference plane T to the lower end of the lens is h mm, so the position of the height reference line Lh can be determined; Make camera 1 reach the height reference position under horizontal conditions; at this time, use a certain characteristic position on camera 1 to record the height difference parameter h1 between camera 1 and the stand; The height difference parameter between the brackets is equal to h1; when h1=h2=h3=h4, it can be considered that the cameras are of the same height and meet the requirements of the lens imaging working distance, the shooting areas of each camera are separated from each other as a whole, and the edge areas are handed over in pairs;

(3) Taking the borders on both sides of the detection surface as the benchmark, mark the detection width range OE of the same height on the camera bracket by the plumb line method. The width of the camera field of view yv for a single camera imaging is w ₁ mm, and the width of the detection area is l _OE length. W is ₂ mm, the field of view of the four cameras of the imaging system completely covers the OE; the positional distance Δd between the cameras is set to the same value, and on the glass substrate detection plane T, the four cameras shoot at three locations in the area. The relationship between the width v of the overlapping area of the field of view and the width of the camera's field of view yv, the width of the overlapping area v, and the distance between the camera installation positions Δd is:

yv=Δd+v (1)

At this time, take O as the origin, O represents the starting point of the bracket, E represents the end point of the bracket, d _i represents the distance from camera i to point O, d _e represents the distance from camera 4 to point E, and Δd represents the difference between camera i and camera i+1 The distance between the points; P1~P4 points respectively represent the positions of the cameras 1~4 projected on the OE, and the _Ly axis coordinate value of the position Pi of the camera i can be expressed as: d _i =d ₁ +(i-1)Δd or d _i =l _OE -d _e -(4-i)Δd, where i = 1, 2, 3, 4; at this time, there is the following relationship between the detection width range of equal height OE and the camera interval Δd:

l _OE =d ₁ +3Δd+d _e (2)

则由公式(2)可得

此时相机1～4在相机支架上的具体位置坐标就可以确定了，之后再按照标记位置平移相机即可；In order to facilitate subsequent image processing, the effective field of view width of each camera should be guaranteed to be equal, and when the effective field of view width of camera 1 and camera 4 are the same, it is obvious that:

Then from formula (2) we can get

At this time, the specific position coordinates of cameras 1 to 4 on the camera bracket can be determined, and then the camera can be translated according to the marked position;

的视场宽度；The width of the overlapping area of the field of view in the detection range: v=w ₁ -Δd, due to the symmetry of the camera imaging system, each camera should be discarded in the left and right areas of the field of view

the width of the field of view;

(2) The front and rear position and parallelism of the camera

(1) Each camera is placed on the camera base for adjusting the camera inclination and pitch angle. Whether the imaging area is parallel to the light reference line is used as the basis for adjusting the parallelism of the camera, and the front and rear positions of the camera are adjusted so that the camera shooting area is within the light range. ;

(2) When the detected object passes through the trigger position S of the photoelectric sensor, the camera starts to capture images. If the inclination angle of the camera is θ at this time, the angle between the straight line l where the shooting area is located and the illumination reference line Ly is also θ; when When the moving distance of the glass substrate is Δs, since the scanning speed of the camera is constant, in the imaging system coordinate system, the captured image area I can be obtained by translating the straight line l, and the translation distance is Δs; The angle between the boundary and the starting boundary of the image is θ; according to the principle of camera imaging, in image I, the angle between the starting boundary of the detected object and the image boundary is also θ; by continuously adjusting the left and right angles of the camera, when:

θ=0 (3)

, the current camera is already parallel to the camera stand. Use the above steps for each camera so that each camera is parallel to the camera stand.

(3) A certain feature point on camera 1 is arbitrarily taken as a reference point, and the distance d ₁ from the camera bracket to the camera bracket is used as the basis to adjust the distance from other cameras to the camera bracket to satisfy:

s ₁ =s ₂ =s ₃ =s ₄ (4)

At this time, the pictures captured by each camera in each frame are in a parallel state.

(3) Camera pitch angle

(1) Taking the imaging brightness of each camera as a reference, when the optical axis of the camera is close to the vertical direction, the shooting area is close to the light reference line, and the image brightness is enhanced. First, use this standard to adjust the camera's pitch angle slightly. The brightness is uniform in the rectangular area. When the camera pitch angle is adjusted to move the camera shooting position within the rectangular area, the images are all highlighted and the brightness changes are not obvious;

(2) Further accurate judgment of the pitch angle by means of motion imaging: by adjusting the pitch angle of camera 1, record the pitch angle range θ ₁ ~ θ _n with high brightness and no obvious brightness change, and take the middle value as the angle of camera 1. value, that is:

θ _m =(θ ₁ +θ _n )/2

The shooting position of camera 1 at this time is regarded as the position of the light reference line; then four cameras are used to image the detected object at the same time through the motion platform; when the detected object passes the trigger position S of the photoelectric sensor, the camera is triggered to start acquisition; the detected object is detected When the object moves Δs, the width x1 of the blank area in the image captured by the camera 1 is used as the benchmark. When the collection position of the camera i is lower than the collection position of the camera 1, the width of the blank area in the image captured by the camera _i > x ₁ , Conversely, when the acquisition position of camera i is higher than the acquisition position of camera 1, x _i <x ₁ ; therefore, when:

x ₁ =x ₂ =x ₃ =x ₄

At the same time, the imaging positions of each camera can be placed on the same line that is very close and parallel to the lighting reference line.

The advantages of the present invention are:

The invention solves the problem that the traditional multi-line array CCD camera coplanar and collinear shooting joint debugging method is difficult to take into account the requirements of calibration efficiency and precision, and realizes high calibration efficiency and high precision.

Description of drawings

Figure 1 is a schematic diagram of the camera height and position problem.

Figure 2 is a schematic diagram of the problem of camera parallel adjustment and front and rear position.

Figure 3 shows the positional relationship between the camera shooting area and the lighting area.

Figure 4 shows the imaging principle of the tilt camera.

Figure 5 is a schematic diagram of the camera pitch angle problem.

FIG. 6 is a schematic diagram of motion imaging in camera pitch adjustment.

Detailed ways

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, but not all of the embodiments.

Example.

A joint debugging method for coplanar and collinear imaging of multi-line CCD cameras. The shooting area of the line CCD camera can be regarded as a fixed-length line segment on the detection reference plane. The exact position, camera height, front-to-back position, parallelism, pitch angle, etc. need to be further adjusted accurately. Therefore, after the lighting system is adjusted, the camera adjustment needs to be performed.

(1) Camera height and camera position

During the installation of the camera, the height of the camera can only be roughly estimated, and the height of each camera is easily adjusted by the naked eye. Since the lens of an industrial camera has a fixed range, when the overall height of the camera is outside the working distance, the camera cannot image clearly; when the height of the camera is uneven, if the camera is completely adjusted to make a clear image, the focal length of the lens is easy to produce big difference. Therefore, the camera height needs to be adjusted and calibrated in advance.

As shown in Fig. 1, in order to describe the spatial position information of the imaging system, it is necessary to establish a spatial coordinate system of the imaging system. Among them, Lx, Ly, and Lz are the three-dimensional reference lines of the imaging system spatial coordinate system, respectively. Lx is the left edge of the TFT-LCD glass substrate of the detection object, Ly is the horizontal reference line drawn when the light source is adjusted, and Lz is a vertical reference line that passes through the intersection of the straight lines Lx and Ly and is perpendicular to the reference plane where the detection object is located. This three-dimensional coordinate system will also be used in subsequent descriptions of the imaging system.

According to the lens documentation, when the standard magnification of the lens is n, it can be calculated that the working distance of the lens is h mm, that is, the distance from the detection reference plane T to the lower end of the lens is h mm. From this, the position of the height reference line Lh can be determined. Adjust the overall camera stand to make the camera 1 reach the height reference position when the stand is horizontal. At this time, a certain characteristic position on the camera 1 is used to record the height difference parameter h1 between the camera 1 and the stand. Taking h1 as the benchmark, adjust the camera base so that the height difference parameter between the other three cameras and the stand is equal to h1. When h1=h2=h3=h4, it can be considered that the camera has the same height and meets the requirements of the lens imaging working distance.

The four cameras of the imaging system need to cooperate with each other to complete the imaging of the entire target surface. Therefore, the shooting areas of each camera should be separated from each other as a whole, and the edge areas should be handed over in pairs. If the camera position is unreasonable, it is easy to have missing areas, and adjusting the camera position in advance is also conducive to the subsequent field of view stitching work.

Based on the borders on both sides of the detection surface, the detection width range OE of the same height is marked on the camera bracket by the plumb line method. It can be seen from the previous section that the width of the camera's field of view for imaging by a single camera is w ₁ mm, the width of the detection area is l, and the length of the _OE is w ₂ mm. Obviously, for the purpose of detection, the field of view of the four cameras should completely cover the OE. In order to facilitate the splicing of the camera's field of view, the position interval Δd between the cameras is preferably set to the same value. When the separation distance Δd of the camera installation positions is the same, it is obvious that on the glass substrate detection plane T, the overlapping area widths v of the three fields of view in the shooting areas of the four cameras are equal, and the camera field of view width yv, the overlapping area width v, the camera installation The relationship between the position interval Δd is:

yv=Δd+v (1)

At this time, take O as the origin, O represents the starting point of the bracket, E represents the end point of the bracket, d _i represents the distance from camera i to point O, d _e represents the distance from camera 4 to point E, and Δd represents the difference between camera i and camera i+1 The distance between the points; P1~P4 points respectively represent the positions of the cameras 1~4 projected on the OE, and the _Ly axis coordinate value of the position Pi of the camera i can be expressed as: d _i =d ₁ +(i-1)Δd or d _i =l _OE -d _e -(4-i)Δd, where i = 1, 2, 3, 4; at this time, there is the following relationship between the detection width range of equal height OE and the camera interval Δd:

l _OE =d ₁ +3Δd+d _e (2)

则由公式(2)可得

此时相机1～4在相机支架上的具体位置坐标就可以确定了，之后再按照标记位置平移相机即可；In order to facilitate subsequent image processing, the effective field of view width of each camera should be guaranteed to be equal, and when the effective field of view width of camera 1 and camera 4 are the same, it is obvious that:

Then from formula (2) we can get

At this time, the specific position coordinates of cameras 1 to 4 on the camera bracket can be determined, and then the camera can be translated according to the marked position;

的视场宽度；The width of the overlapping area of the field of view in the detection range: v=w ₁ -Δd, due to the symmetry of the camera imaging system, each camera should be discarded in the left and right areas of the field of view

the width of the field of view;

(2) The front and rear position and parallelism of the camera

When a machine vision inspection system uses an industrial camera, the camera is generally installed on a camera base that can be adjusted to various angles, and then the base is installed on the camera bracket. The camera base used in this article to adjust the camera tilt and pitch angle is a conforming camera base composed of two layers of precision angular displacement stages.

During the installation process, it is not completely guaranteed that the front and rear positions of each camera are aligned, nor can it be judged whether the cameras are parallel to the lighted area. When the camera is not installed in parallel, since the camera and its shooting area are strictly parallel, the camera shooting area will be affected and inclined at a certain angle to the lighting area. If no adjustment is made at this time, the four cameras will not be able to reach the same shooting baseline (lighting baseline), as shown in Figure 2:

In the figure, l′ _i (i=1, 2, 3, 4) is a straight line formed by translating the straight line where the camera shooting area is located to the coordinate origin. A slight unevenness in the camera installation process will prevent the camera shooting area from being perfectly matched to the lighting area.

As shown in Figure 3, when the front and rear position of the camera is close to the appropriate position, and the camera is tilted left and right, the situation described in Figures a and b will occur. When the camera pitch problem is not considered, the situation described in Figures c and d will occur when the camera is too far forward or backward. And the imaging method described in Figure e is what we finally need to achieve.

When solving the problem of camera non-parallel and front and rear, we should first solve the problem of camera non-parallel. When the camera is in a parallel state, we can randomly select a feature point on the camera, and use the horizontal distance between the point and the bracket to replace the front and rear position coordinates of the camera as a reference for adjusting the front and rear position of the camera.

Since the camera is strictly parallel to its imaging area, whether the imaging area is parallel to the light reference line can be used as the basis for adjusting the parallelism of the camera. In order to make the image of the camera clear and the brightness is appropriate, the front and rear positions of the camera can be adjusted so that the shooting area of the camera can be within the illumination range. Due to the problem of the pitch angle, the front and rear positions of the camera cannot be used as the final result at this time, but need to be further adjusted by the calibration method described above after the camera parallel problem is solved.

In order to adjust the parallelism of the camera, the imaging process is shown in Figure 4(a). When the detected object passes the trigger position S of the photoelectric sensor, the camera starts to capture images. If the tilt angle of the camera is θ at this time, the angle between the straight line l where the shooting area is located and the lighting reference line Ly is also θ. When the moving distance of the glass substrate is Δs, since the scanning speed of the camera is constant, in the imaging system coordinate system, the captured image area I can be obtained by translating the straight line l, and the translation distance is Δs. Obviously, in the coordinate system of the imaging system, the angle between the boundary of the detected object and the starting boundary of the image is θ. According to the camera imaging principle, in image I, the angle between the starting boundary of the detected object and the image boundary is also θ; by continuously adjusting the left and right angles of the camera, when:

θ=0 (3)

, the current camera is already parallel to the camera stand. Use the above steps for each camera so that each camera is parallel to the camera stand.

Although the camera pitch angle has not been adjusted at this time, the pitch angle of each camera is relatively small, so the influence on the front and rear positions of the feature points on the camera is very small and can be ignored. Therefore, a certain feature point on the camera 1 can be arbitrarily taken as a reference point, and the distance d ₁ from the camera bracket to the camera bracket can be used as the basis to adjust the distance from other cameras to the camera bracket to satisfy:

s ₁ =s ₂ =s ₃ =s ₄ (4)

At this time, the pictures captured by each camera in each frame are in a parallel state.

(3) Camera pitch angle

The camera is successively adjusted in height in the Lz direction, in the installation position in the Ly direction, and in the front and rear positions in the Lx direction, and make the camera parallel to the light reference line. At this time, as shown in Figure 5(b), theoretically, on the plane where the imaging systems Lx and Lz are located, the imaging positions of the four cameras will coincide at a certain point on the vertical reference line Lz. Due to the problem of camera pitch angle, the optical axis of the camera and the vertical reference line Lz will form a very small angle. The shooting situation of the camera at this time is shown in Fig. 5(a). Obviously, we can take the imaging brightness of each camera as a reference. When the optical axis of the camera is close to the vertical direction, the shooting area is close to the lighting baseline, and the image brightness is enhanced. This standard can only be used to adjust the tilt angle of the camera slightly. Since the illumination area of the strip light source has uniform brightness in a rectangular area, when the tilt angle of the camera is adjusted to move the camera shooting position within the rectangular area, the images are all highlighted. And the brightness change is not obvious.

At this time, the pitch angle should be further accurately judged by means of motion imaging. Before that, we can record the pitch angle range θ ₁ ~ θ _n with high brightness and no obvious brightness change by adjusting the pitch angle of camera 1, and take the middle value as the accurate angle value of camera 1, namely:

θ _m =(θ ₁ +θ _n )/2

The shooting position of the camera 1 at this time is regarded as the position of the light reference line. Then, four cameras are used to image the detected object at the same time through the motion platform. Its imaging effect is shown in Figure 6. When the detected object passes the trigger position S of the photoelectric sensor, the camera is triggered to start collecting. When the detected object moves Δs, take the blank area width x1 in the image captured by camera 1 as the benchmark, when the capture position of camera i is lower than the capture position of camera 1, the blank area width x _i > x in the captured image ₁ , on the contrary, when the acquisition position of camera i is higher than the acquisition position of camera 1 , x _i <x ₁ . Therefore, when:

x ₁ =x ₂ =x ₃ =x ₄

At the same time, the imaging positions of each camera can be placed on the same line that is very close and parallel to the lighting reference line.

The above description is only a preferred embodiment of the present invention, but the protection scope of the present invention is not limited to this. The equivalent replacement or change of the inventive concept thereof shall be included within the protection scope of the present invention.

Claims (1)

1. a multi-line array CCD camera coplanar collinear imaging joint debugging method, is characterized in that, comprises the following steps: (1) First, adjust the camera height and camera position; (2) Readjust the front and rear position and parallelism of the camera; (3) Finally adjust the camera pitch angle; The adjustment of the camera height and camera position in step (1) specifically includes the following steps: (1) Establish the imaging system spatial coordinate system to describe the spatial position information of the imaging system. Let Lx, Ly, and Lz be the three-dimensional reference lines of the imaging system spatial coordinate system respectively; Lx is the left edge of the detection object TFT-LCD glass substrate , Ly is the horizontal reference line drawn when the light source is adjusted, Lz is a vertical reference line that passes through the intersection of the straight lines Lx and Ly and is perpendicular to the reference plane where the detection object is located; (2) When the standard magnification of the lens is n, the working distance of the lens is h mm, that is, the distance from the detection reference plane T to the lower end of the lens is h mm, so the position of the height reference line Lh can be determined; Make camera 1 reach the height reference position under horizontal conditions; at this time, use a certain characteristic position on camera 1 to record the height difference parameter h1 between camera 1 and the stand; The height difference parameter between the brackets is equal to h1; when h1=h2=h3=h4, it can be considered that the cameras are of the same height and meet the requirements of the lens imaging working distance, the shooting areas of each camera are separated from each other as a whole, and the edge areas are handed over in pairs; (3) Taking the borders on both sides of the detection surface as the benchmark, mark the detection width range OE of the same height on the camera bracket by the plumb line method. The width of the camera field of view yv for a single camera imaging is w ₁ mm, and the width of the detection area is l _OE length. W is ₂ mm, the field of view of the four cameras of the imaging system completely covers the OE; the positional distance Δd between the cameras is set to the same value, and on the glass substrate detection plane T, the four cameras shoot at three locations in the area. The relationship between the width v of the overlapping area of the field of view and the width of the camera's field of view yv, the width of the overlapping area v, and the distance between the camera installation positions Δd is: yv=Δd+v (1) At this time, O represents the starting point of the bracket, E represents the end point of the bracket, d _i represents the distance from camera i to point O, d _e represents the distance from camera 4 to point E, Δd represents the distance between camera i and camera i+1; P1 The ~P4 points represent the positions of the cameras 1 to 4 projected onto the OE, respectively, and the _Ly -axis coordinate value of the position Pi of the camera i can be expressed as: d _i =d ₁ +(i-1)Δd or d _i =l _OE − d _e -(4-i)Δd, where i=1, 2, 3, 4; at this time, there is the following relationship between the detection width range of equal height OE and the camera interval Δd: l _OE =d ₁ +3Δd+d _e (2) In order to facilitate subsequent image processing, the effective field of view width of each camera should be guaranteed to be equal, and when the effective field of view width of camera 1 and camera 4 are the same, it is obvious that:

Then from formula (2) we can get

At this time, the specific position coordinates of cameras 1 to 4 on the camera bracket can be determined, and then the camera can be translated according to the marked position; The width of the overlapping area of the field of view in the detection range: v=w ₁ -Δd, due to the symmetry of the camera imaging system, each camera should be discarded in the left and right areas of the field of view

the width of the field of view; The problem of adjusting the front and rear position and parallelism of the camera in step (2) specifically includes the following steps: (1) Each camera is placed on the camera base for adjusting the camera inclination and pitch angle. Whether the imaging area is parallel to the light reference line is used as the basis for adjusting the parallelism of the camera, and the front and rear positions of the camera are adjusted so that the camera shooting area is within the light range. ; (2) When the detected object passes through the trigger position S of the photoelectric sensor, the camera starts to capture images. If the inclination angle of the camera is θ at this time, the angle between the straight line l where the shooting area is located and the illumination reference line Ly is also θ; when When the moving distance of the glass substrate is Δs, since the scanning speed of the camera is constant, in the imaging system coordinate system, the captured image area I can be obtained by translating the straight line l, and the translation distance is Δs; The angle between the boundary and the starting boundary of the image is θ; according to the principle of camera imaging, in image I, the angle between the starting boundary of the detected object and the image boundary is also θ; by continuously adjusting the left and right angles of the camera, when: θ=0 (3) , the current camera is already parallel to the camera bracket; apply the above steps to each camera so that each camera is parallel to the camera bracket; (3) Arbitrarily take a certain feature point on camera 1 as a reference point, and adjust the distance from other cameras to the camera bracket based on the distance s ₁ from it to the camera bracket to satisfy: s ₁ =s ₂ =s ₃ =s ₄ (4) At this time, the pictures taken by each camera in each frame are in a parallel state; The step (3) of adjusting the camera pitch angle specifically includes the following steps: (1) Taking the imaging brightness of each camera as a reference, when the optical axis of the camera is close to the vertical direction, the shooting area is close to the light reference line, and the image brightness is enhanced. First, use this standard to adjust the camera's pitch angle slightly. The brightness is uniform in the rectangular area. When the camera pitch angle is adjusted to move the camera shooting position within the rectangular area, the images are all highlighted and the brightness changes are not obvious; (2) Further accurate judgment of the pitch angle by means of motion imaging: by adjusting the pitch angle of camera 1, record the pitch angle range θ ₁ ~ θ _n with high brightness and no obvious brightness change, and take the middle value as the angle of camera 1. value, that is: θ _m =(θ ₁ +θ _n )/2 The shooting position of camera 1 at this time is regarded as the position of the light reference line; then four cameras are used to image the detected object at the same time through the motion platform; when the detected object passes the trigger position S of the photoelectric sensor, the camera is triggered to start acquisition; the detected object is detected When the object moves Δs, the width x1 of the blank area in the image captured by the camera 1 is used as the benchmark. When the collection position of the camera i is lower than the collection position of the camera 1, the width of the blank area in the image captured by the camera _i > x ₁ , Conversely, when the acquisition position of camera i is higher than the acquisition position of camera 1, x _i <x ₁ ; therefore, when: x ₁ =x ₂ =x ₃ =x ₄ At the same time, the imaging positions of each camera can be placed on the same line that is very close and parallel to the lighting reference line.

Images