CN114782553B - Iterative Camera Calibration Method and Device Based on Ellipse Dual Conic - Google Patents

Abstract

The invention relates to an iterative camera calibration method based on an elliptic dual conic, which comprises the steps of extracting the center coordinates of an ellipse on a calibration plate image; calibrating a camera according to two-dimensional pixel coordinates and three-dimensional space coordinates corresponding to the circle center coordinates of the ellipse, and solving to obtain camera parameters and a first re-projection error; correcting perspective projection and lens distortion in the calibration plate image, and converting the calibration plate image into a forward view; extracting the center coordinates of the perfect circle on the corrected calibration plate image; re-projecting the circle center coordinates of the perfect circle back to a camera coordinate system, calibrating the camera again according to the coordinates of the circle center coordinates of the perfect circle under the camera coordinate system, and solving to obtain a second re-projection error; and carrying out convergence judgment according to the first re-projection error and the second re-projection error. The invention provides an iterative camera calibration method based on an elliptic dual conic, which can reduce the reprojection error by 92.4%, and greatly improve the calibration precision of a camera.

Description

Iterative Camera Calibration Method and Device Based on Ellipse Dual Conic

technical field

The invention relates to the technical field of machine vision, in particular to an iterative camera calibration method and device based on an ellipse dual quadratic curve.

Background technique

In the process of image measurement and machine vision applications, in order to determine the relationship between the three-dimensional geometric position of a point on the surface of a space object and its corresponding point in the image, it is necessary to establish a geometric model of camera imaging. These geometric model parameters are camera parameters. Under most conditions, these parameters must be obtained through experiments and calculations. This process of solving parameters is called camera calibration. Therefore, in machine vision applications, the calibration of camera parameters is a very critical link. However, in machine vision, Because the corners of the checkerboard are sensitive to noise and image quality, the detection accuracy of the corner coordinates is low, and the circular features have a strong suppression of noise and high detection accuracy. are widely used in.

However, when the camera is calibrated with the circular mode planar calibration plate, it mainly relies on a set of contour points extracted in advance. Since it cannot perform multiple calibrations, it is very sensitive to non-uniform illumination such as light sources and reflections. The projected ellipse The center of the circle is not necessarily the projection of the center of the perfect circle. At the same time, the ring is also affected by distortion, and then eccentricity error occurs, which affects the camera calibration accuracy.

Contents of the invention

For this reason, the technical problem to be solved by the present invention is to overcome the problems existing in the prior art, and propose an iterative camera calibration method and device based on the elliptic dual conic curve, which can reduce the re-projection error by 92.4%, greatly improving the camera calibration accuracy.

In order to solve the above-mentioned technical problems, the present invention provides an iterative camera calibration method based on ellipse dual conic curve, comprising the following steps:

S1: Use the camera to take multiple shots of the calibration plate to obtain an image of the calibration plate, and preprocess the image of the calibration plate;

S2: Using the ellipse dual quadratic curve to extract the coordinates of the center of the ellipse on the calibration plate image;

S3: Calibrate the camera according to the two-dimensional pixel coordinates and three-dimensional space coordinates corresponding to the coordinates of the center of the ellipse, and obtain the initial values of the camera parameters and the first reprojection error;

S4: Correct the perspective projection and lens distortion in the calibration plate image according to the initial value of the camera parameters, convert the calibration plate image to the front view, correct the projected ellipse to an approximate perfect circle, and obtain the corrected calibration plate image;

S5: Using the ellipse dual quadratic curve to extract the coordinates of the center of the perfect circle on the calibration plate image after correction;

S6: According to the initial value of the camera parameters, re-project the coordinates of the center of the perfect circle back to the original camera coordinate system, re-calibrate the camera according to the coordinates of the center of the perfect circle in the camera coordinate system, and solve to obtain the second re-projection error;

S7: Perform convergence judgment according to the first re-projection error and the second re-projection error, if the second re-projection error is smaller than the first re-projection error, then repeat the operations from S4 to S7, if the second re-projection error is greater than or equal to the first re-projection error If there is no reprojection error, the operation ends.

In one embodiment of the present invention, in S2, the method for preprocessing the calibration plate image includes:

S2.1: Define the Gaussian function, use the Gaussian filter to calculate the image gradient in the area containing the ellipse in the calibration plate image, the image gradient defines the normal direction passing through the pixel center, where the Gaussian function is defined as In the formula, σ is the standard deviation, u is the mean;

S2.2: Convert the Gaussian function into a 5×5 filter template, set the convolution kernel scale factor to 1, and calculate the horizontal and vertical gradients of each pixel in the calibration plate image as G(x) and G(y ), the gradient direction, that is, the normal direction, is defined as

S2.3: Define the weight of each tangent line perpendicular to the normal direction, and obtain the set of tangent lines, where the weight formula is defined as

S2.4: Use the tangent set to obtain the parameters of the ellipse dual quadratic curve, and obtain the coordinates of the center of the ellipse by solving the parameters of the ellipse dual quadratic curve.

In one embodiment of the present invention, in S2.4, the method of obtaining the ellipse dual conic parameters by using the tangent set, and obtaining the coordinates of the center of the ellipse by solving the ellipse dual conic parameters includes:

S2.41: The perspective projection of the space circle onto the imaging plane is a quadratic curve, the equation of which the center of the circle is at the origin is Ax ² +Bxy+Cy ² +Dx+Ey+F=0(4), where, A, B, C, D, E, F are the coefficients of the quadratic curve, the tangent in the tangent set is called the dual quadratic curve, and its parameter Q ^* is defined as In the formula, A', B', C', D', E', F' are the coefficients of the dual conic curve:

S2.42: Given a set of tangent lines l _i , the parameter vector of the dual quadratic curve Q ^* is set to ψ={A',B',C',D',E',F'}, Q ^* passes linear The least squares estimation makes the approximate fitting function Φ(ψ) reach the minimum value, and obtains the parameters of the dual quadratic curve, where In the formula, l _i is a set of tangent lines, l _i ^T is the transpose of the tangent line set matrix, ω is the weight of each tangent line, and ψ is the parameter vector of the dual quadratic curve Q ^* ;

S2.43: Invert the dual quadratic curve coefficient matrix to obtain the elliptic curve coefficient matrix, and obtain the coordinates of the center of the ellipse through formula (7) and formula (8) and /> In the formula, x ₀ and y ₀ are the coordinates of the center of the ellipse on the x and y axes respectively, and A, B, C, D, E, and F are the coefficients of the quadratic curve.

In one embodiment of the present invention, in S3, the method for solving the camera parameters and the first reprojection error includes:

S3.1: Calculate the initial value of the camera parameters in the ideal case of no distortion, and consider the lens distortion at the same time, and use the least square method to calculate the nonlinear distortion parameters;

S3.2: Calculate the first re-projection error according to the internal and external parameters of the camera and nonlinear distortion parameters In the formula, n is the number of ellipse center points in the calibration plate image, m _j is the real coordinate of the jth ellipse center point in the calibration plate image, /> is the reprojection coordinates of the center point of the jth ellipse in the calibration plate image, (u ₀ , v ₀ ) is the coordinates of the origin after translation, f _x and f _y represent the scale factors of the horizontal and vertical directions of the image respectively, k ₁ , k ₂ , k ₃ , k ₄ are the radial distortion coefficients of the camera, p ₁ , p ₂ are the tangential distortion coefficients of the camera, and RMS is the first reprojection error.

In one embodiment of the present invention, in step S3.1, in the ideal case of no distortion, the camera model is a small hole imaging model, and the space coordinates of the center of the circle are set to P(X _W , Y _W , Z _W ) , the process of projecting the point to a point p(u,v) on the two-dimensional pixel coordinate system is as follows:

In the formula, (u ₀ , v ₀ ) is the coordinates of the origin, f _x , f _y are the focal lengths of the camera in the x and y directions respectively, A is the internal reference matrix, and (RT) is the external reference matrix.

In one embodiment of the present invention, in step S3.1, the nonlinear camera model involves the radial distortion of the camera as follows:

In the formula, (x _u , y _u ) is the image coordinate after distortion, (x, y) is the ideal undistorted image coordinate, and k ₁ , k ₂ , k ₃ , k ₄ , p ₁ and p ₂ are all distortion factor, is the radial distortion, is the tangential distortion.

In one embodiment of the present invention, in S4, the method for converting the calibration plate image to the front view includes:

The image of the calibration plate is converted from the viewing plane with an oblique angle to the positive viewing plane through reverse perspective transformation, and the matrix H is used as the product of the internal and external parameter matrices in formula (10), and the matrix H is The matrix H is the perspective transformation matrix Put formula (12) into formula (10) to get Use the formula (14) to eliminate the scale factor Z _c to obtain the conversion relationship between the perspective projected coordinate point and the two-bit pixel coordinate point/> and /> Invert the matrix H to get /> Thus, the coordinates (u', v') of each point in the front view can be obtained, namely and />

In one embodiment of the present invention, in S6, the method of reprojecting the coordinates of the center of the perfect circle back to the original camera coordinate system includes:

Reproject the coordinates of the center of the perfect circle back to the camera coordinate system by the following transformation formula:

In the formula, (u', v') are the undistorted coordinates of each pixel in the two-dimensional pixel coordinate system in the front view, H ₁₁ ', H ₁₂ ', H ₁₃ ', H ₂₁ ', H ₂₂ ', H ₂₃ ′, H ₃₁ ′, H ₃₂ ′, H ₃₃ ′ are parameters of the anti-perspective projection matrix.

In addition, the present invention also provides a computer device, including a memory, a processor, and a computer program stored on the memory and operable on the processor, and the processor implements the steps of the above-mentioned method when executing the program.

Moreover, the present invention also provides a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, the steps of the above-mentioned method are realized.

The above technical solution of the present invention has the following advantages compared with the prior art:

The iterative camera calibration method based on the elliptic dual quadratic curve proposed by the present invention uses the parameters obtained by the traditional calibration method as initialization parameters, reprojects the calibration image to the front view for correction, and then obtains the approximate standard on the front view The coordinates of the calibration points of the circle are re-calibrated using the coordinates of the calibration points of the perfect circle, and the eccentricity error is reduced through multiple iterations to obtain a higher-precision calibration point, which can reduce the re-projection error by 92.4%, which greatly improves the camera. Calibration accuracy.

Description of drawings

In order to make the content of the present invention more clearly understood, the present invention will be further described in detail below according to the specific embodiments of the present invention and in conjunction with the accompanying drawings.

Fig. 1 is the flowchart of camera calibration method in the present invention;

Fig. 2 is the used ring calibration plate of image acquisition among the present invention;

Fig. 3 is the result after extracting the center of circle based on the dual quadratic curve in the present invention;

Fig. 4 is the calibration plate after preprocessing and anti-perspective transformation in the present invention;

Fig. 5 is a reprojection error distribution diagram of the traditional calibration method;

Fig. 6 is a reprojection error distribution diagram of the calibration method proposed in the present invention;

Fig. 7 is an iterative diagram of the reprojection error of the calibration method proposed in the present invention.

Detailed ways

The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments, so that those skilled in the art can better understand the present invention and implement it, but the examples given are not intended to limit the present invention.

Please refer to FIG. 1 to FIG. 4, the embodiment of the present invention provides an iterative camera calibration method based on ellipse dual conic curve, including the following steps:

S1: Use the camera to take multiple shots of the calibration plate to obtain an image of the calibration plate, and preprocess the image of the calibration plate;

S2: Using the ellipse dual quadratic curve to extract the coordinates of the center of the ellipse on the calibration plate image;

S3: Calibrate the camera according to the two-dimensional pixel coordinates and three-dimensional space coordinates corresponding to the coordinates of the center of the ellipse, and obtain the initial values of the camera parameters and the first reprojection error;

S4: Correct the perspective projection and lens distortion in the calibration plate image according to the initial value of the camera parameters, convert the calibration plate image to the front view, correct the projected ellipse to an approximate perfect circle, and obtain the corrected calibration plate image;

S5: Using the ellipse dual quadratic curve to extract the coordinates of the center of the perfect circle on the calibration plate image after correction;

S6: According to the initial value of the camera parameters, re-project the coordinates of the center of the perfect circle back to the original camera coordinate system, re-calibrate the camera according to the coordinates of the center of the perfect circle in the camera coordinate system, and solve to obtain the second re-projection error;

S7: Perform convergence judgment according to the first re-projection error and the second re-projection error, if the second re-projection error is smaller than the first re-projection error, then repeat the operations from S4 to S7, if the second re-projection error is greater than or equal to the first re-projection error If there is no reprojection error, the operation ends.

As a specific embodiment of the present application, the specific operation of preprocessing the calibration plate image in step S2 includes the following steps:

S2.1: Define the Gaussian function, use the Gaussian filter to calculate the image gradient in the area containing the ellipse in the calibration plate image, the image gradient defines the normal direction passing through the pixel center, where the Gaussian function is defined as:

In the formula, σ is the standard deviation, u is the mean, and the mean of the Gaussian function is set to 0;

S2.2: Convert the Gaussian function into a 5×5 filter template, set the convolution kernel scale factor to 1, and calculate the horizontal and vertical gradients of each pixel in the calibration plate image as G(x) and G(y ), then the gradient direction, that is, the direction of the normal line, can be defined as;

S2.3: Define the weight of each tangent perpendicular to the normal direction, focus on the active area of the gradient, and give higher attention to the place where the gradient is stronger, thereby obtaining a set of tangents, where the weight formula is defined as :

S2.4: Use the tangent set to obtain the parameters of the ellipse dual quadratic curve, and obtain the coordinates of the center of the ellipse by solving the parameters of the ellipse dual quadratic curve.

In the above step S2.4, the perspective projection of the space circle onto the imaging plane is a quadratic curve, assuming that its center of circle is at the origin and its equation is:

Ax ² +Bxy+Cy ² +Dx+Ey+F＝0 (4)

In the formula, A, B, C, D, E, F are the coefficients of the quadratic curve. The relationship of the plane to the conic is represented using a series of straight lines tangent to the conic, these tangents are called dual conic, whose parameter Q ^* is defined as:

Given a set of tangents l _i , the parameter vector of the dual quadratic curve Q ^* is set to ψ={A',B',C',D',E',F'}, Q ^* can be obtained by linear least squares Estimation, so that the approximate fitting function Φ(ψ) reaches the minimum value, and then the parameters of the dual quadratic curve can be obtained;

In the formula, l _i is the set of tangent lines, l _i ^T is the transposition of the set matrix of tangent lines, ω is the weight of each tangent line, and ψ is the parameter vector of the dual quadratic curve Q ^* ;

The coefficient matrix of the dual quadratic curve is inverted to obtain the coefficient matrix of the elliptic curve, and the coordinates of the center of the ellipse are obtained by the following formula:

In the formula, x ₀ and y ₀ are the coordinates of the center of the ellipse on the x and y axes respectively, and A, B, C, D, E, and F are the coefficients of the quadratic curve.

As a specific embodiment of the present application, before step S2.1, the calibration plate image is first converted into a grayscale image; then the grayscale image is binarized using the maximum inter-class difference method; finally, the grayscale image is removed by dilation and corrosion. Irrelevant information in the degree image, only the most essential ellipse information is retained.

As a specific embodiment of the present application, the specific operation of step S3 includes the following steps:

S3.1: Calculate the initial value of the camera parameters in the ideal case of no distortion, and consider the lens distortion at the same time, and use the least square method to calculate the nonlinear distortion parameters;

S3.2: According to the internal and external parameters of the camera and the nonlinear distortion parameters, the first re-projection error is calculated as follows:

In the formula, n is the number of ellipse center points in the calibration board image, _mj is the real coordinate of the jth ellipse center point in the calibration board image, is the reprojection coordinates of the center point of the jth ellipse in the calibration plate image, (u ₀ , v ₀ ) is the coordinates of the origin after translation, f _x and f _y represent the scale factors of the horizontal and vertical directions of the image respectively, k ₁ , k ₂ , k ₃ , k ₄ are the radial distortion coefficients of the camera, p ₁ , p ₂ are the tangential distortion coefficients of the camera, and RMS is the first reprojection error, which is an index for evaluating the calibration accuracy of the camera;

S3.3: Use the reprojection error obtained in S3.2 as the objective function, and use the maximum likelihood estimation method to optimize the initial value of the camera parameters.

As a specific embodiment of the present application, in step S3.1, in the ideal case of no distortion, the camera model is a small hole imaging model, and the spatial coordinates of the center of the ring are set to P(X _W , Y _W , Z _W ), The process of projecting the point to a point p(u,v) on the two-dimensional pixel coordinate system is as follows:

In the formula, (u ₀ , v ₀ ) is the coordinates of the origin; f _x , f _y are the focal lengths of the camera in the x and y directions respectively, and the unit is pixel; A is the internal reference matrix; (RT) is the external reference matrix.

As a specific embodiment of the present application, in step S3.1, the nonlinear camera model involves the radial distortion of the camera as follows:

In the formula, (x _u , y _u ) are the distorted image coordinates, (x, y) are the ideal undistorted image coordinates, k ₁ , k ₂ , k ₃ , k ₄ , p ₁ and p ₂ are all distortion factor, is the radial distortion, is the tangential distortion.

As a specific embodiment of the present application, in step S4, the essence of converting the original calibration plate image to the forward view is to convert the image from a viewing plane with an oblique angle to a forward viewing plane through inverse perspective transformation . In order to simplify the calculation, the world coordinate system is fixed on the target plane, then the physical coordinate Z _w of any point on the target plane = 0, and the three-dimensional world coordinates and the two-dimensional pixel coordinates of each image are normalized, as follows :

Let the matrix H be the product of the internal and external parameter matrices in formula (10):

The matrix H is the homography matrix, that is, the perspective transformation matrix:

Put formula (12) into formula (10) to get:

Using formula (14) to eliminate the scale factor Z _c , we can get:

So far, the transformation relationship between the perspective projected coordinate point and the two-bit pixel coordinate point has been obtained, and the inversion of the matrix H can be obtained:

The coordinates (u ^' , v') of each point in the front view can be obtained;

As a specific embodiment of the present application, in step S6, since the perspective projection matrix of the known camera is formula (13), (u', v') are the coordinates of each point in the front view, (x', y') are the coordinates of each point of (u', v') in the original camera coordinate system, and the coordinates of the center of the perfect circle can be reprojected back to the camera coordinate system by the following transformation formula.

As a specific embodiment of the present application, in step S7, whether the number of iterations exceeds 10 or whether the reprojection error in formula (9) decreases generation by generation is used as the basis for judging convergence.

The iterative camera calibration method based on the elliptic dual quadratic curve proposed by the present invention uses the parameters obtained by the traditional calibration method as initialization parameters, reprojects the calibration image to the front view for correction, and then obtains the approximate standard on the front view The coordinates of the calibration points of the circle are re-calibrated using the coordinates of the calibration points of the perfect circle, and the eccentricity error is reduced through multiple iterations to obtain a higher-precision calibration point, which can reduce the re-projection error by 92.4%, which greatly improves the camera. Calibration accuracy.

In order to verify the performance of the present invention, the method of the present invention and the traditional method are used to calibrate the camera as a comparison. Specifically, for the 13 captured images, the camera is calibrated respectively by the method of the present invention and the traditional method, and the obtained results are shown in Table 1 below.

Table 1 Comparison of camera calibration results

In camera calibration research, the reprojection error is usually used to judge the camera calibration accuracy. The re-projection error refers to the re-projection of the internal and external parameters of the camera and the distortion parameters obtained by calibration to the three-dimensional point in space, and the deviation between the coordinates of the new projection point of the three-dimensional point on the image and the coordinates of the original imaging point. Generally, the smaller the reprojection error, the higher the accuracy of camera calibration. The projection errors of the method of the present invention and the traditional method for all images are shown in Table 2 below. It can be seen from Table 2 that the overall average error of the method of the present invention is 0.0671, which is 92.4% lower than that of the traditional method. In the traditional method, the reprojection error distribution of all feature points in each image is shown in Figure 5, and the error points are seriously diffused. The reprojection error distribution of all feature points in each image by the method of the present invention is shown in Figure 6, and the error points are basically concentrated within 0.2. The combination of Table 2, Figure 5 and Figure 6 shows that the method in this paper greatly improves the camera calibration accuracy, and also verifies the feasibility and effectiveness of the method of the present invention. It is worth mentioning that the method of the present invention has a fast convergence speed, and the convergence target can be reached after about 3 iterations, as shown in FIG. 7 .

Table 2 Overall average reprojection error comparison of camera calibration Unit: pixel

Corresponding to the above method embodiment, the embodiment of the present invention also provides a computer device, including:

memory for storing computer programs;

The processor is used for implementing the steps of the above-mentioned iterative camera calibration method based on the elliptic dual conic when executing the computer program.

In the embodiment of the present invention, the processor may be a central processing unit (Central Processing Unit, CPU), an application-specific integrated circuit, a digital signal processor, a field programmable gate array, or other programmable logic devices.

The processor can call the program stored in the memory, specifically, the processor can execute the operations in the embodiment of the method for quickly calculating three-dimensional polarization dimensions.

The memory is used to store one or more programs, the programs may include program codes, and the program codes include computer operation instructions.

In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as at least one magnetic disk storage device or other volatile solid-state storage devices.

Corresponding to the above method embodiment, the embodiment of the present invention also provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the above-mentioned method based on the elliptic dual conic curve is realized. Steps for an iterative camera calibration method.

Those skilled in the art should understand that the embodiments of the present application may be provided as methods, systems, or computer program products. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) having computer-usable program code embodied therein.

The present application is described with reference to flowcharts and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the present application. It should be understood that each procedure and/or block in the flowchart and/or block diagram, and a combination of procedures and/or blocks in the flowchart and/or block diagram can be realized by computer program instructions. These computer program instructions may be provided to a general purpose computer, special purpose computer, embedded processor, or processor of other programmable data processing equipment to produce a machine such that the instructions executed by the processor of the computer or other programmable data processing equipment produce a An apparatus for realizing the functions specified in one or more procedures of the flowchart and/or one or more blocks of the block diagram.

These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing apparatus to operate in a specific manner, such that the instructions stored in the computer-readable memory produce an article of manufacture comprising instruction means, the instructions The device realizes the function specified in one or more procedures of the flowchart and/or one or more blocks of the block diagram.

These computer program instructions can also be loaded onto a computer or other programmable data processing device, causing a series of operational steps to be performed on the computer or other programmable device to produce a computer-implemented process, thereby The instructions provide steps for implementing the functions specified in the flow chart or blocks of the flowchart and/or the block or blocks of the block diagrams.

Apparently, the above-mentioned embodiments are only examples for clear description, and are not intended to limit the implementation. For those of ordinary skill in the art, on the basis of the above description, other changes or changes in various forms can also be made. It is not necessary and impossible to exhaustively list all the implementation manners here. However, the obvious changes or changes derived therefrom are still within the scope of protection of the present invention.

Claims (10)

1. an iterative camera calibration method based on the ellipse dual quadratic curve, is characterized in that, comprises the following steps: S1: Use the camera to take multiple shots of the calibration plate to obtain an image of the calibration plate, and preprocess the image of the calibration plate; S2: Using the ellipse dual quadratic curve to extract the coordinates of the center of the ellipse on the calibration plate image; S3: Calibrate the camera according to the two-dimensional pixel coordinates and three-dimensional space coordinates corresponding to the coordinates of the center of the ellipse, and obtain the initial values of the camera parameters and the first reprojection error; S4: Correct the perspective projection and lens distortion in the calibration plate image according to the initial value of the camera parameters, convert the calibration plate image to the front view, correct the projected ellipse to an approximate perfect circle, and obtain the corrected calibration plate image; S5: Using the ellipse dual quadratic curve to extract the coordinates of the center of the perfect circle on the calibration plate image after correction; S6: According to the initial value of the camera parameters, re-project the coordinates of the center of the perfect circle back to the original camera coordinate system, re-calibrate the camera according to the coordinates of the center of the perfect circle in the camera coordinate system, and solve to obtain the second re-projection error; S7: Perform convergence judgment according to the first re-projection error and the second re-projection error, if the second re-projection error is smaller than the first re-projection error, then repeat the operations from S4 to S7, if the second re-projection error is greater than or equal to the first re-projection error If there is no reprojection error, the operation ends. 2. the iterative camera calibration method based on elliptic dual conic according to claim 1, is characterized in that, in S1, the method for preprocessing the calibration plate image comprises: S1.1: Define the Gaussian function, use the Gaussian filter to calculate the image gradient in the area containing the ellipse in the calibration plate image, the image gradient defines the normal direction passing through the pixel center, where the Gaussian function is defined as In the formula, σ is the standard deviation, u is the mean; S1.2: Convert the Gaussian function into a 5×5 filter template, set the convolution kernel scale factor to 1, and calculate the horizontal and vertical gradients of each pixel in the calibration plate image as G(x) and G(y ), the gradient direction, that is, the normal direction, is defined as S1.3: Define the weight of each tangent perpendicular to the normal direction, and obtain the set of tangents, where the weight formula is defined as S1.4: Obtain the ellipse dual conic parameters by using the tangent set, and obtain the coordinates of the ellipse center by solving the ellipse dual conic parameters. 3. The iterative camera calibration method based on the ellipse dual quadratic curve according to claim 2, characterized in that in S1.4, the ellipse dual quadratic curve parameters are obtained by using the tangent set, and according to the ellipse dual quadratic curve parameters The methods for obtaining the coordinates of the center of the ellipse include: S1.41: The perspective projection of the space circle onto the imaging plane is a quadratic curve, the equation of which the center of the circle is at the origin is Ax ² +Bxy+Cy ² +Dx+Ey+F=0(4), where, A, B, C, D, E, F are the coefficients of the quadratic curve, the tangent in the tangent set is called the dual quadratic curve, and the matrix Q ^* is defined as In the formula, A', B', C', D', E', F' are the coefficients of the dual conic curve: S1.42: Given a set of tangent lines l _i , the parameter vector of Q ^* is set to ψ={A',B',C',D',E',F'}, and Q ^* is estimated by linear least squares Make the approximate fitting function Φ(ψ) reach the minimum value, and obtain the parameters of the dual quadratic curve, where In the formula, l _i is the set of tangent lines, l _i ^T is the transpose of the set matrix of tangent lines, ω is the weight of each tangent line, and ψ is the parameter vector of Q ^* ; S1.43: Invert the dual quadratic curve coefficient matrix to obtain the elliptic curve coefficient matrix, and obtain the coordinates of the center of the ellipse through formula (7) and formula (8) and /> In the formula, x ₀ and y ₀ are the coordinates of the center of the ellipse on the x and y axes respectively, and A, B, C, D, E, and F are the coefficients of the quadratic curve. 4. the iterative camera calibration method based on elliptic dual conic according to claim 1, is characterized in that, in S3, the method for obtaining the initial value of camera parameter and the first reprojection error for solving comprises: S3.1: Calculate the initial value of the camera parameters in the ideal case of no distortion, and consider the lens distortion at the same time, and use the least square method to calculate the nonlinear distortion parameters; S3.2: Calculate the first re-projection error according to the internal and external parameters of the camera and nonlinear distortion parameters In the formula, n is the number of ellipse center points in the calibration plate image, m _j is the real coordinate of the jth ellipse center point in the calibration plate image, /> is the reprojection coordinates of the center point of the jth ellipse in the calibration plate image, (u ₀ , v ₀ ) is the coordinates of the origin after translation, f _x and f _y are the focal lengths of the camera in the x and y directions respectively, k ₁ , k ₂ , k ₃ , k ₄ are the radial distortion coefficients of the camera, p ₁ , p ₂ are the tangential distortion coefficients of the camera, and RMS is the first reprojection error. 5. The iterative camera calibration method based on the elliptic dual quadratic curve according to claim 4, characterized in that, in step S3.1, in the ideal case of no distortion, the camera model is a pinhole imaging model, and the ring The spatial coordinates of the center of the circle are P(X _W , Y _W , Z _W ), and the process of projecting the center of the circle to a point p(u,v) on the two-dimensional pixel coordinate system is as follows: In the formula, (u ₀ , v ₀ ) are the coordinates of the origin after translation, f _x and f _y are the focal lengths of the camera in the x and y directions respectively, A is the internal reference matrix, (RT) is the external reference matrix, and Z _c is the scale factor. 6. The iterative camera calibration method based on the elliptic dual quadratic curve according to claim 4, characterized in that: in step S3.1, the nonlinear camera model involves the radial distortion of the camera as follows: In the formula, (x _u , y _u ) is the image coordinate after distortion, (x, y) is the ideal undistorted image coordinate, and k ₁ , k ₂ , k ₃ , k ₄ , p ₁ and p ₂ are all distortion factor, is the radial distortion, /> is the tangential distortion. 7. the iterative camera calibration method based on elliptic dual conic according to claim 5, is characterized in that, in S4, the method for converting the calibration plate image to the front view comprises: The image of the calibration plate is converted from the viewing plane with an oblique angle to the positive viewing plane through reverse perspective transformation, and the matrix H is used as the product of the internal and external parameter matrices in formula (10), and the matrix H is The matrix H is the perspective projection matrix Put formula (12) into formula (10) to get Use the formula (14) to eliminate the scale factor Z _c to obtain the conversion relationship between the perspective projected coordinate point and the two-bit pixel coordinate point/> and /> Invert the matrix H to get /> Thus, the coordinates (u', v') of each point in the front view can be obtained, namely and /> 8. The iterative camera calibration method based on the ellipse dual conic according to claim 7, wherein in S6, the method of reprojecting the coordinates of the center of a perfect circle back to the original camera coordinate system comprises: Reproject the coordinates of the center of the perfect circle back to the camera coordinate system by the following transformation formula: In the formula, (u', v') are the undistorted coordinates of each pixel in the two-dimensional pixel coordinate system in the front view, H ₁₁ ', H ₁₂ ', H ₁₃ ', H ₂₁ ', H ₂₂ ', H ₂₃ ′, H ₃₁ ′, H ₃₂ ′, H ₃₃ ′ are parameters of the anti-perspective projection matrix. 9. A computer device, comprising a memory, a processor, and a computer program stored on the memory and operable on the processor, characterized in that, when the processor executes the program, any one of claims 1 to 8 is realized The steps of the method. 10. A computer-readable storage medium, on which a computer program is stored, characterized in that, when the program is executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.