A Fast Calibration Method and System for the Intrinsic Parameters of an On-vehicle Fish-eye Camera

Abstract

The present invention belongs to the technical field of in-vehicle camera calibration, and specifically relates to a method and system for quickly calibrating the internal parameters of an in-vehicle fisheye camera, including: obtaining the fixed internal parameters, distortion coefficients, error ranges of the internal parameters, and corner coordinates of a standard calibration cloth when the camera leaves the factory; performing distortion correction on the corner coordinates of the calibration cloth captured by the camera according to the distortion coefficients and assuming the internal parameters of the camera; calculating the homography matrix between the corrected corner coordinates and the corner coordinates of the standard calibration cloth; performing inverse calculation using the homography matrix to obtain the inverse-pushed corner coordinates of the calibration cloth; calculating the standard deviation between the inverse-pushed corner coordinates of the calibration cloth and the corner coordinates of the standard calibration cloth; and obtaining the optimal internal parameter value corresponding to the minimum standard deviation through polling calculation. The present invention calibrates the internal parameters of the camera once by using the calibration cloth to obtain a more accurate internal parameter value.

Description

Technical Field

[0001] The present invention belongs to the technical field of in-vehicle camera calibration, and particularly relates to a method and system for quickly calibrating the internal parameters of an in-vehicle fisheye camera. Background Technique

[0002] In order to obtain a larger viewing angle, in-vehicle surround-view cameras generally use fisheye wide-angle lenses. However, due to production process and other issues, during the production of fisheye cameras, the position of the optical center often shifts and is generally not at the center point of the image output by the camera, as Figure 1 shown.

[0003] Camera manufacturers often provide a fixed internal parameter value of the camera. If this fixed internal parameter value is used for calibration, the obtained results are often not ideal, as Figure 2 shown. It can be seen that there is obvious misalignment in the splicing of the checkerboard.

[0004] Due to the error in the optical center of the camera at the factory, if the internal parameters of the camera are not optimized and corrected, the distortion correction result may be in error. The traditional internal parameters are calculated by Zhang Zhengdingyou's checkerboard calibration method, which requires multiple checkerboard photos at different positions and angles to calculate the camera internal parameters. However, this method has limitations and is not applicable to product off-line. Summary of the Invention

[0005] Object of the Invention: Due to the production process problems of fisheye cameras, there will be a certain error in their optical centers, and this error will also have a certain impact on the subsequent external parameter calibration, thereby affecting the overall calibration result. The present invention provides a method for quickly calibrating the internal parameters of an in-vehicle fisheye camera, which can obtain a more accurate internal parameter value by using a calibration cloth to calibrate the internal parameters of the camera once.

[0006] Technical Solution: The method for quickly calibrating the internal parameters of the in-vehicle fisheye camera according to the present invention includes the following steps:

[0007] S1. Obtain the fixed internal parameters and distortion coefficients of the camera at the factory, as well as the standard calibration cloth corner point coordinates of the arranged calibration cloth;

[0008] S2. Obtain the error range of the internal parameters of the camera at the factory and , where is the error range of the pixel scale, the error range of the coordinates at the line of sight focus;

[0009] S3. Obtain the corner point coordinates of the calibration cloth by shooting with the camera;

[0010] S4. Assume that the form of the internal parameters of the camera is as follows:

[0011] ,

[0012] Among them, is the coordinate at the focus of the camera's line of sight in the image coordinate system, is the pixel scale. It is assumed that in the camera internal parameters the initial value is , the initial value of remains unchanged;

[0013] S5. According to the distortion coefficients, the assumed camera internal parameters, and the coordinates of the calibration cloth corner points obtained from the camera shooting, use the OpenCV library function fisheye::undistortPoints() to perform distortion correction to obtain the corrected corner point coordinates;

[0014] S6. Calculate the homography matrix between the corrected corner point coordinates and the standard calibration cloth corner point coordinates;

[0015] S7. Perform back-calculation based on the corrected corner point coordinates and the homography matrix to obtain the back-calculated calibration cloth corner point coordinates;

[0016] S8. Calculate the standard deviation between the back-calculated calibration cloth corner point coordinates and the standard calibration cloth corner point coordinates;

[0017] S9. Repeat S5 - S8. Through polling calculation, make the coordinate at the focus of the line of sight increase from to . The corresponding when the standard deviation is the smallest during the process is the of the optimal internal parameters;

[0018] S10. Assume that in the camera internal parameters the initial value is , the initial value of remains unchanged. Repeat S5 - S8. Through polling calculation, make the pixel scale increase from to . The corresponding when the standard deviation is the smallest during the process is the of the optimal internal parameters;

[0019] S11. Combine and output the obtained in S9 and the obtained in S10 as the optimal internal parameters.

[0020] To further improve the above technical solution, the calibration cloth is composed of four checkerboards and four rectangular blocks. The four checkerboards are distributed in the front, rear, and both sides of the vehicle, and the four rectangular blocks are distributed at the four corners of the vehicle and are aligned with the checkerboards in the horizontal and vertical directions; the corner points of the calibration cloth include the corner points of the checkerboards and the four corner points of the rectangular blocks.

[0021] Further, the corner coordinates of the calibration cloth captured by the camera include: capturing an image of the calibration cloth by the camera, and using OpenCV to extract the corner coordinates (eye_point) of the calibration cloth on the captured image in the image coordinate system.

[0022] Further, the distortion correction is performed using the fisheye::undistortPoints() function provided by OpenCV. The input parameters of the function include: assuming the camera internal parameters (intrisic), distortion coefficients (coeffs), and the corner coordinates of the calibration cloth captured by the camera (eye_point). The function output is the corrected corner coordinates (out_point_line).

[0023] Further, using the findHomography() function in OpenCV, the corrected corner coordinates (out_point_line) are matched with the known standard calibration cloth corner coordinates (pts_board) to obtain a homography matrix (m_find_homography_vec) that describes the perspective transformation between the image coordinate system and the standard calibration cloth coordinate system for subsequent back-calculation.

[0024] Further, applying the homography matrix (m_find_homography_vec) to the corrected corner coordinates (out_point_line), performing a linear combination on each point of the corrected corner coordinates (out_point_line), and inversely mapping to obtain the corresponding coordinates in the standard calibration cloth coordinate system. The correctness of the inverse mapping result is ensured through a normalization operation, and the back-calculated calibration cloth corner coordinates are saved in the back-calculated calibration cloth corner coordinates (homo_get_point).

[0025] Further, the standard deviation between the back-calculated calibration cloth corner coordinates (homo_get_point) and the corner coordinates of the standard calibration cloth (pts_board) includes: traversing all corner pairs between the back-calculated calibration cloth corner coordinates (homo_get_point) and the corner coordinates of the standard calibration cloth (pts_board), calculating the squared difference between each corner pair and accumulating them, and evaluating the accuracy of the currently assumed camera internal parameters through the sum of the returned squared differences.

[0026] A system for implementing the above method for quickly calibrating the internal parameters of a vehicle-mounted fisheye camera includes:

[0027] A camera parameter acquisition module for acquiring the fixed internal parameters, distortion coefficients, and error range of the internal parameters of the camera when it leaves the factory and ;

[0028] A corner point coordinate extraction module, which is used to obtain the calibration cloth image captured by the camera and extract the calibration cloth corner point coordinates in the calibration cloth image;

[0029] A hypothetical camera internal parameter setting module, assuming that the form of the camera internal parameters is as follows:

[0030] ,

[0031] where, is the coordinate at the focus of the camera's line of sight in the image coordinate system, is the pixel scale, and the setting rule is: assume that the initial value in the camera internal parameters is , remains unchanged, the scale is 1, so that the coordinate at the focus of the line of sight increases from to , in the camera internal parameters the initial value is , the initial value of remains unchanged, the scale is 1, so that the pixel scale increases from to , and output the hypothetical camera internal parameters according to the setting rule;

[0032] A coordinate correction module, which is used to obtain the hypothetical camera internal parameters and distortion coefficients, and perform distortion correction on the extracted calibration cloth corner point coordinates to obtain the corrected corner point coordinates;

[0033] A homography matrix calculation module, which is used to calculate the homography matrix between the corrected corner point coordinates and the standard calibration cloth corner point coordinates;

[0034] A corner point calculation module, which is used to perform inverse calculation on the corrected corner point coordinates according to the homography matrix to obtain the coordinates of the inverse-projected calibration cloth corner points;

[0035] A standard deviation calculation module, which is used to calculate the standard deviation between the coordinates of the inverse-projected calibration cloth corner points and the standard calibration cloth corner point coordinates;

[0036] A polling calculation module, which is used to poll the hypothetical camera internal parameter values output by the hypothetical camera internal parameter setting module, and according to the hypothetical camera internal parameter values, use the coordinate correction module, homography matrix calculation module, corner point calculation module, and standard deviation calculation module to obtain the standard deviation, and determine the optimal internal parameter value when the standard deviation is the smallest.

[0037] Beneficial effects: Compared with the prior art, the advantages of the present invention are as follows: A method for calibrating the internal parameters of a camera when an automobile comes off the production line is provided herein. Based on the parameters of the camera when it leaves the factory, first, a photo of the calibration cloth checkerboard is taken by the camera to obtain the coordinates of the calibration cloth, and the coordinates after undistortion are obtained through inverse distortion. Then, the homography matrix is calculated from the coordinates of the calibration cloth and the coordinates of the standard calibration cloth. Using the obtained homography matrix and the undistorted coordinates obtained above, new coordinates of the calibration cloth are obtained. After obtaining these coordinates of the calibration cloth, they are compared with the coordinates of the standard calibration cloth to obtain the minimum error, and finally the optimal solution is obtained.

[0038] The main advantages of this method are short calibration time, the ability to quickly calibrate to obtain accurate internal parameters, directly using the original calibration site without the need to add other calibration materials, saving costs. Moreover, the internal parameters of the calibrated camera are more accurate, which plays an important role in subsequent external parameter calibration, making image stitching more accurate and having fewer ghost images in the stitching area. Brief Description of the Drawings

[0039] Figure 1 is the rendering taken by the fish-eye camera;

[0040] Figure 2 is the rendering of calibration using the internal parameters of the camera when it leaves the factory;

[0041] Figure 3 is the layout diagram of the standard calibration cloth;

[0042] Figure 4 is the flowchart of the method provided by the present invention;

[0043] Figure 5 is the rendering of calibration after optimizing the internal parameters by the present invention. Detailed Description of the Embodiment

[0044] The technical solution of the present invention will be described in detail below with reference to the drawings, but the protection scope of the present invention is not limited to the described embodiments.

[0045] Embodiment 1: Based on the parameters of the camera when it leaves the factory as a benchmark, first, the fish-eye coordinates of the calibration cloth are obtained using OpenCV image recognition, and the coordinates after undistortion are obtained through inverse distortion. Then, the homography matrix is calculated from the coordinates of the calibration cloth and the coordinates of the standard calibration cloth. Using the obtained homography matrix and the undistorted coordinates obtained above, new coordinates of the calibration cloth are obtained. After obtaining these coordinates of the calibration cloth, they are compared with the coordinates of the standard calibration cloth to obtain the minimum error, and finally the optimal solution is obtained. As Figure 3As shown, the calibration cloth consists of four checkerboards and four rectangular blocks. The four checkerboards are distributed in the front, rear, and both sides of the vehicle, and the four rectangular blocks are distributed at the four corners of the vehicle and are aligned with the checkerboards in the horizontal and vertical directions. The corner points of the calibration cloth include the corner points of the checkerboards and the four corner points of the rectangular blocks.

[0046] As Figure 4 shown in the on-vehicle fish-eye camera internal parameter rapid calibration method, it includes the following steps:

[0047] S1. Obtain the fixed internal parameters, distortion coefficients, and the corner point coordinates of the standard calibration cloth when the camera leaves the factory.

[0048] S2. Obtain the error range of the internal parameters when the camera leaves the factory.

[0049] S3. Obtain the corner point coordinates of the calibration cloth captured by the camera.

[0050] S4. Assume the form of the camera internal parameters is as follows:

[0051] ,

[0052] where is the position of the focus of the camera's line of sight in the image coordinate system. Assume that the initial value in the camera internal parameters takes the lower limit according to the error range of the internal parameters in S2. is the pixel scale. Assume that the initial value in the camera internal parameters is determined according to the value in the fixed internal parameters.

[0053] S5. According to the distortion coefficients, the assumed camera internal parameters, and the corner point coordinates of the calibration cloth captured by the camera, use the OpenCV library function fisheye::undistortPoints() for distortion correction to obtain the corrected corner point coordinates.

[0054] S6. Calculate the homography matrix between the corrected corner point coordinates and the corner point coordinates of the standard calibration cloth.

[0055] S7. Perform back-calculation according to the corrected corner point coordinates and the homography matrix to obtain the back-calculated calibration cloth corner point coordinates.

[0056] S8. Calculate the standard deviation between the back-calculated calibration cloth corner point coordinates and the corner point coordinates of the standard calibration cloth.

[0057] S9. Repeat S4 - S9. Through polling calculation, make the coordinate at the focus of the line of sight increase from to . The corresponding to the minimum standard deviation during the process is the of the optimal internal parameters.

[0058] S10. Assume that in the internal parameters of the camera, the initial value is , and the initial value of remains unchanged. Repeat S5 - S8. Through polling calculation, the pixel scale is increased from to . The corresponding to the minimum standard deviation during the process is the

[0059] optimal internal parameter of ; S11. Combine and output the

[0060] obtained in S9 and the Figure 5 obtained in S10 as the optimal internal parameter.

[0061] Example 2: A system for implementing the method for quickly calibrating the internal parameters of an in - vehicle fish - eye camera in Example 1, including:

[0062] A camera parameter acquisition module, used to acquire the fixed internal parameters, distortion coefficients, and the error range of the internal parameters when the camera leaves the factory;

[0063] A corner coordinate extraction module, used to acquire the fish - eye image captured by the camera and extract the corner coordinates of the calibration cloth in the fish - eye image;

[0064] A fish - eye image coordinate correction module, used to acquire the internal parameters of the camera and the distortion coefficients, correct the distortion of the extracted corner coordinates of the calibration cloth, and obtain the corrected corner coordinates. Assume the form of the internal parameters of the camera is as follows:

[0065] ,

[0066] where is the position of the focus of the camera's line of sight in the image coordinate system. Assume that in the internal parameters of the camera, the initial value takes the lower limit according to the error range of the internal parameters, is the pixel scale. Assume that in the internal parameters of the camera, the initial value is determined according to the value in the fixed internal parameters;

[0067] A homography matrix calculation module, used to calculate the homography matrix between the corrected corner coordinates and the standard calibration cloth corner coordinates;

[0068] A corner calculation module, used to perform inverse - push calculation on the corrected corner coordinates according to the homography matrix to obtain the coordinates of the inverse - pushed calibration cloth corners;

[0069] The standard deviation calculation module is used to calculate the standard deviation between the coordinates of the back-calculated calibration cloth corner points and the coordinates of the standard calibration cloth corner points;

[0070] The polling calculation module is used to adjust the internal camera parameters according to the standard deviation calculation result to obtain the optimal internal parameters.

[0071] Example 3: According to the corner point coordinates of the calibration cloth obtained by the camera shooting, perform distortion correction on the corner point coordinates.

[0072] 1. Distortion correction

[0073] First, call the OpenCV function fisheye::undistortPoints() to perform distortion correction on the corner points on the captured fisheye image. Part of the code is as follows:

[0074] std::vector <cv::point2f>out_point_line;

[0075] fisheye::undistortPoints(eye_point, out_point_line, intrisic, coeffs);

[0076] In the above code: eye_point is the original corner point coordinates obtained by the camera. Due to the distortion of the fisheye lens, these points are usually stretched or distorted, resulting in inaccurate positions; out_point_line stores the corner point coordinates after distortion correction; intrisic and coeffs are the internal parameters of the camera (including focal length and optical center) and the lens distortion coefficients respectively. These parameters are used to describe the characteristics of the fisheye lens and help correct the distortion.

[0077] Through the fisheye::undistortPoints() function, the original corner point coordinates are mapped to a corrected coordinate system, thus eliminating the influence of distortion.

[0078] 2. Coordinate scale transformation

[0079] In the corrected coordinate system, the x and y coordinates will undergo a certain scale adjustment and are transformed in combination with the internal parameters of the camera (normal_fx, normal_fy, normal_cx, normal_cy).

[0080] ,

[0081] normal_fx, normal_fy: represent the scale factors of the focal length and are used to adjust the scale of the coordinates.

[0082] normal_cx, normal_cy: represent the coordinates of the optical center position and are used to translate the origin of the coordinate system to the actual optical center position of the camera.

[0083] This transformation adjusts the original distortion-corrected coordinates to coordinates that conform to the standard pixel coordinate system. The purpose of this processing is to be able to compare and verify with the corner point coordinates of the known standard calibration cloth.

[0084] The above obtains the coordinates of the points after distortion correction.

[0085] 3. Calculate the homography matrix

[0086] Next, through the corner points after distortion correction and the corner point coordinates of the standard calibration cloth, use the findHomography() function to calculate the homography matrix between the two sets of coordinates.

[0087] homography_camera = findHomography(out_point_line, pts_board);

[0088] m_find_homography_vec = homography_camera.reshape(1, 1);

[0089] findHomography(): This function calculates a homography matrix that describes the perspective transformation from one coordinate system to another. Here, the two coordinate systems are the corner coordinates after distortion correction and the standard corner coordinates on the calibration board.

[0090] homography_camera: Stores the calculated homography matrix.

[0091] m_find_homography_vec: The homography matrix is expanded into a one-dimensional vector for subsequent coordinate transformation.

[0092] The homography matrix reflects the geometric relationship between the distortion correction result of the camera and the actual calibration board.

[0093] For the obtained homography matrix, based on the obtained distortion-corrected points, the coordinate points of the calculated calibration board are inversely deduced.

[0094] 4. Inverse Coordinate Deduction and Calculation

[0095] Through the calculated homography matrix, the distortion-corrected points are further inversely mapped to the coordinate system of the standard calibration board to verify whether the position of each point is accurate.

[0096] ,

[0097] Inverse Transformation: According to the elements in the homography matrix, the corrected corner coordinates are converted back to the corresponding coordinates on the calibration board.

[0098] point_z: The normalization factor for the projective transformation to ensure the accurate perspective effect of the coordinates.

[0099] tmp_point.x, tmp_point.y: The coordinate points of the calibration board obtained by inverse deduction.

[0100] 5. Calculation of Standard Deviation and Optimization

[0101] By calculating the distance difference between the coordinate points of the calibration board obtained by inverse deduction and the actual coordinates of the standard calibration board, the accuracy of the internal parameters is evaluated. The final result is to verify the effects of distortion correction and homography matrix calculation by comparing the sum of the squared differences (standard deviation) of all points.

[0102] Then, the sum of the squared differences is obtained by using the points obtained above and the coordinate points of the standard calibration cloth.

[0103] ;

[0104] Calculation of squared differences: The error of each point is obtained by calculating the differences (distance_x and distance_y) between the corner points obtained by back-calculation and the corner points of the standard calibration cloth.

[0105] Return the standard deviation: Return the sum of the squared errors, which is used as an index to evaluate the optimization of the model.

[0106] Finally, its standard deviation is obtained. Poll the above steps, compare each parameter of the internal parameters, and finally obtain the optimal solution.

[0107] As described above, although the present invention has been shown and described with reference to specific preferred embodiments, it should not be construed as a limitation of the present invention itself. Various changes in form and detail may be made without departing from the spirit and scope of the present invention defined by the appended claims.

Claims

1. A method for quickly calibrating the internal parameters of a vehicle-mounted fish-eye camera, characterized in that, It includes the following steps: S1. Obtain the fixed internal parameters and distortion coefficients of the camera at the time of factory production, as well as the standard calibration cloth corner point coordinates of the arranged calibration cloth; S2. Obtain the error range of the internal parameters of the camera at the time of factory production and , is the error range of the pixel scale, the error range of the coordinates at the line of sight focus; S3. Obtain the corner point coordinates of the calibration cloth by shooting with the camera; S4. Assume that the form of the camera internal parameters is as follows: , Among them, is the coordinate at the focus of the camera's line of sight in the image coordinate system, is the pixel scale; S5. According to the distortion coefficients, the assumed camera internal parameters, and the corner point coordinates of the calibration cloth obtained by the camera shooting, use the OpenCV library function fisheye::undistortPoints() to perform distortion correction to obtain the corrected corner point coordinates; S6. Calculate the homography matrix between the corrected corner point coordinates and the standard calibration cloth corner point coordinates; S7. Perform back-propagation calculation based on the corrected corner point coordinates and the homography matrix to obtain the back-propagated calibration cloth corner point coordinates; S8. Calculate the standard deviation between the back-propagated calibration cloth corner point coordinates and the standard calibration cloth corner point coordinates; S9. Assume that in the camera internal parameters The initial value is , The initial value of remains unchanged. Repeat S5 - S8. Through polling calculation, the coordinates at the line of sight focus increase from to . The corresponding when the standard deviation is the smallest during the process is the of the optimal internal parameters; S10. Assume that in the internal parameters of the camera the initial value is , and the initial value of remains unchanged. Repeat S5 - S8. Through polling calculation, the pixel scale is increased from to . The corresponding to the minimum standard deviation during the process is the of the optimal internal parameters; S11. Combine the obtained in S9 with the obtained in S10 and output them as the optimal internal reference.

2. The method for quickly calibrating the internal parameters of an in-vehicle fish-eye camera according to claim 1, characterized in that, The calibration cloth is composed of four checkerboards and four rectangular blocks. The four checkerboards are distributed in the front, rear, and both sides of the vehicle, and the four rectangular blocks are distributed at the four corners of the vehicle and are aligned with the checkerboards in the horizontal and vertical directions; the corner points of the calibration cloth include the corner points of the checkerboards and the four corner points of the rectangular blocks.

3. The method for quickly calibrating the internal parameters of an in-vehicle fish-eye camera according to claim 2, wherein The corner point coordinates of the calibration cloth obtained by the camera shooting include: shooting the image of the calibration cloth by the camera, and using OpenCV to extract the corner point coordinates (eye_point) of the calibration cloth on the captured image in the image coordinate system.

4. The method for quickly calibrating the internal parameters of an in-vehicle fisheye camera according to claim 3, wherein, The distortion correction is performed using the fisheye::undistortPoints() function provided by OpenCV. The input parameters of the function include: the assumed camera internal parameters (intrisic), the distortion coefficients (coeffs), and the corner point coordinates of the calibration cloth obtained by the camera shooting (eye_point). The function output is the corrected corner point coordinates (out_point_line).

5. The method for quickly calibrating the internal parameters of an in-vehicle fish-eye camera according to claim 4, wherein Use the findHomography() function in OpenCV to match the corrected corner point coordinates (out_point_line) with the known standard calibration cloth corner point coordinates (pts_board) to obtain the homography matrix (m_find_homography_vec) describing the perspective transformation between the image coordinate system and the standard calibration cloth coordinate system for subsequent back-propagation calculation.

6. The method for quickly calibrating the internal parameters of an in-vehicle fisheye camera according to claim 5, characterized in that, Apply the homography matrix (m_find_homography_vec) to the corrected corner point coordinates (out_point_line), perform a linear combination on each point of the corrected corner point coordinates (out_point_line), and reverse-map to obtain the corresponding coordinates in the standard calibration cloth coordinate system. Ensure the correctness of the reverse-mapping result through normalization operation, and save the back-propagated calibration cloth corner point coordinates to the back-propagated calibration cloth corner point coordinates (homo_get_point).

7. The method for quickly calibrating the internal parameters of an in-vehicle fish-eye camera according to claim 5, characterized in that, The standard deviation between the corner coordinates of the reverse projection calibration cloth (homo_get_point) and the corner coordinates of the standard calibration cloth (pts_board) includes: traversing all corner pairs between the corner coordinates of the reverse projection calibration cloth (homo_get_point) and the corner coordinates of the standard calibration cloth (pts_board), calculating the squared difference between each corner pair and accumulating them, and evaluating the accuracy of the currently assumed camera internal parameters through the sum of the returned squared differences.

8. A system for implementing the method for quickly calibrating the internal parameters of the in-vehicle fish-eye camera according to claim 1, characterized in that, The system includes: A camera parameter acquisition module, which is used to acquire the fixed internal parameters, distortion coefficients and error ranges of the internal parameters of the camera at the time of factory and ; A corner coordinate extraction module, configured to obtain a calibration cloth image captured by a camera and extract the calibration cloth corner coordinates in the calibration cloth image; An assumed camera internal parameter setting module, assuming the form of the camera internal parameters as follows: , Among them, is the coordinate at the focus of the camera's line of sight in the image coordinate system, is the pixel scale, and the setting rule is: assuming that the initial value in the camera internal parameters is , remains unchanged, the scale is 1, so that the coordinate at the focus of the line of sight increases from to . Among the camera internal parameters, the initial value is , the initial value remains unchanged, the scale is 1, so that the pixel scale increases from to , and output the assumed camera internal parameters according to the setting rule; A coordinate correction module, configured to obtain the assumed camera internal parameters and distortion coefficients, and perform distortion correction on the extracted calibration cloth corner coordinates to obtain the corrected corner coordinates; A homography matrix calculation module, configured to calculate the homography matrix between the corrected corner coordinates and the corner coordinates of the standard calibration cloth; A corner calculation module, configured to perform reverse calculation on the corrected corner coordinates according to the homography matrix to obtain the coordinates of the reverse projection calibration cloth corners; A standard deviation calculation module, configured to calculate the standard deviation between the coordinates of the reverse projection calibration cloth corners and the corner coordinates of the standard calibration cloth; A polling calculation module, configured to poll the assumed camera internal parameter values output by the assumed camera internal parameter setting module, and according to the assumed camera internal parameter values, use the coordinate correction module, the homography matrix calculation module, the corner calculation module, and the standard deviation calculation module to obtain the standard deviation, and determine the optimal internal parameter value when the standard deviation is the smallest.

Images