CN117611684A - Structural parameter optimization calibration method for biprism virtual binocular vision system - Google Patents

Abstract

The invention relates to a structural parameter optimization calibration method for a biprism virtual binocular vision system, which comprises the following steps: s1: adjusting the relative position of the camera and the biprism; s2: acquiring three-dimensional images of a plurality of groups of calibration plates; s3: acquiring initial values of internal parameters and structural parameters of a virtual binocular camera; s4: acquiring a stereoscopic image sequence, and establishing a space measurement field consisting of characteristic points with regular distribution; s5: performing distortion correction on sub-images corresponding to the left virtual camera and the right virtual camera by using the internal parameters of the virtual binocular camera obtained in the step S3; s6: acquiring feature point matching point pairs; s7: measuring three-dimensional coordinates of feature points in a field; s8: and constructing an objective function based on the space distance error and epipolar constraint, performing nonlinear iterative optimization on the structural parameters, and updating the three-dimensional coordinates of the feature points in the step S7 by using the structural parameters obtained by each iterative calculation. The method can meet the comprehensive requirements of the biprism virtual binocular system in terms of three-dimensional measurement and two-dimensional epipolar geometric constraint.

Description

A structural parameter optimization and calibration method for biprism virtual binocular vision system

Technical field

The invention belongs to the technical field of image digital processing and relates to a visual calibration method, specifically a structural parameter optimization calibration method for a biprism virtual binocular vision system.

Background technique

The virtual binocular vision system based on the principle of biprism refraction is widely used in arc molten pool morphology measurement, stereomicroscope and other fields because of its compact structure and no binocular synchronization imaging error. The stereoscopic image refracted by the biprism and acquired by a single camera can be equivalent to a pair of binocular stereoscopic vision images captured by two virtual cameras, and is called a virtual binocular vision system. The imaging principle of this system is shown in Figure 1.

As we all know, high-precision calibration is the key to three-dimensional vision measurement. According to previous research, it is feasible and effective to apply the traditional binocular calibration method to the biprism virtual binocular system. The traditional dual-target calibration method generally uses two cameras to acquire the same target image and extract feature points, obtain the initial values of the internal participation structure parameters through a linear method, and finally perform nonlinear optimization of the initial calibration values using error constraints. Research in this field generally focuses on introducing different error constraints and using different optimization objective functions to perform nonlinear optimization calibration of parameters. In addition, it is ideal to regard the single-lens stereo vision system based on biprism as a virtual binocular system. For example, the virtual binocular system does not have a physical camera principal point. Therefore, directly applying traditional calibration methods to a virtual binocular system will limit its measurement accuracy accordingly.

Compared with the traditional binocular stereo vision system, the biprism virtual binocular vision system is more suitable for measurement work within a certain field of view, and the relative position of the measured object and the visual system is relatively fixed. For example, through search, it was found that the Chinese patent with announcement number CN108830906B provides an automatic calibration method of camera parameters based on the principle of virtual binocular vision. The device includes a pan-tilt rotating bracket, a camera, an external trigger circuit, a PC, and a chessboard with black and white grids. The calibration plate of the pattern and two plane mirrors arranged at an angle; the two plane mirrors are arranged behind the pan/tilt rotating bracket; the calibration plate is placed at the pan/tilt rotating bracket; the camera is arranged in front of the pan/tilt rotating bracket and the camera side faces the two plane mirrors; when the camera When the calibration board is photographed, the image of the calibration board it captures includes the left and right virtual images formed by the calibration board in the two plane mirrors; when the calibration board rotates an angle, the external trigger circuit controls the camera to take pictures of the calibration board; the PC captures the image of the calibration board. The calibration plate image is processed to obtain and calibrate the internal parameters, distortion parameters and external parameters of the camera in the world coordinate system.

The Chinese patent with publication number CN114111637A provides a three-dimensional reconstruction method based on virtual binocular striped structured light, including constructing a virtual binocular striped structured light vision system based on a biprism; calibrating the virtual binocular vision system to obtain a virtual binocular The internal and external parameters of the two cameras are projected; a three-frequency four-step sinusoidal stripe image is projected to the object; the unfolded phase corresponding to the left and right virtual cameras is obtained using the multi-frequency heterodyne unwrapping method; the unfolded phase is polarized according to the internal and external parameters of the virtual camera. Line correction; perform stereo matching on the left and right unfolded phases to obtain the parallax of the matching points corresponding to the left and right unfolded phases; combine the internal and external parameters of the virtual binoculars and use the parallax principle to convert the parallax into depth information; generate dense points of the object based on the three-dimensional spatial coordinates of the object Cloud image to complete the three-dimensional reconstruction of objects. The above two invention patents aim to replace the traditional binocular system with a virtual binocular vision system to calibrate parameters, but the calibration method is too cumbersome. Another example is CN107121109A, a structured light parameter calibration device and method based on a front-coated plane mirror. , including: placing the front-coated plane mirror and the plane glass target in front of the camera, the camera simultaneously shooting the plane glass target image and its mirror image, establishing a virtual binocular measurement model, and obtaining the front-coated plane mirror coordinate system to the camera coordinates through nonlinear optimization methods is the optimized solution of the rotation matrix and translation vector. The least squares method is used to calculate the image elimination point of the candidate feature points. The white printing paper is placed in front of the front coated plane mirror, and the camera simultaneously captures the actual light strip image and the mirrored light strip image. , extract the center point of the light strip image, calculate the three-dimensional coordinates of the matching point, and solve the light plane equation. The present invention improves the quality of the light strip, improves the accuracy of extracting the center of the light strip, and provides feature points with micron-level position accuracy, obtains a greater number of calibration points, has higher calibration accuracy, and the calibration results are more stable. The calibration object is not a biprism or a prism. Virtual binocular vision systems with similar structures have different algorithm logic involved in calibration. Therefore, inventing a biprism virtual binocular calibration method based on the characteristics of the biprism virtual binocular system plays a key role in improving the three-dimensional measurement accuracy of the system.

Contents of the invention

Based on the shortcomings of the existing technology, the purpose of the present invention is to provide a structural parameter optimization and calibration method for a biprism virtual binocular vision system to improve the three-dimensional measurement accuracy of the system.

The present invention solves its technical problems by adopting the following technical solutions:

A structural parameter optimization and calibration method for biprism virtual binocular vision system,

S1: Adjust the relative position of the camera and the biprism so that the biprism virtual binocular vision system obtains evenly divided left and right sub-images;

S2: Place the calibration plate at different positions under the depth of field range of the biprism virtual binocular vision system, and obtain three-dimensional images of multiple sets of calibration plates;

S3: Evenly segment the original stereo image of the calibration plate to obtain two sub-images with the same size and extract the feature points on the left and right sub-images respectively. Perform rough calibration of the biprism virtual binocular vision system to obtain the virtual binocular Initial values of the camera’s intrinsic structural parameters;

S4: Place the measurement board in front of the vision system, move the measurement board multiple times at a fixed distance and obtain a stereoscopic image sequence, and establish a spatial measurement field composed of regularly distributed feature points;

S5: Evenly segment several sets of stereo images in the spatial measurement field, and use the internal parameters of the virtual binocular camera obtained in step S3 to perform distortion correction on the sub-images corresponding to the left and right virtual cameras;

S6: Extract the center feature points on the distortion-corrected measurement plate image, match the center feature points of the left and right sub-images according to the order of the circle center feature point sequences, and obtain feature point matching point pairs;

S7: Combined with the intrinsic structural parameters of the virtual binocular camera, the three-dimensional coordinates of the feature points in the spatial measurement field are restored through linear triangulation;

S8: Construct an objective function based on spatial distance error and epipolar constraints, perform nonlinear iterative optimization of structural parameters, and the structural parameters calculated in each iteration will be used to update the three-dimensional coordinates of the feature points in step S7.

Moreover, in the biprism virtual binocular vision system, the stereoscopic image refracted by the biprism and acquired by a single camera can be equivalent to a binocular stereoscopic image pair captured by two virtual cameras.

Moreover, the specific steps of S1 are: using a precision shift stage to adjust the positions of the camera and the biprism so that the center line in the vertical direction of the image closely coincides with the edge of the center of the biprism.

Moreover, the biprism virtual binocular vision system was roughly calibrated using Zhang Zhengyou's calibration method.

Moreover, the specific steps of S4 are: fix a measurement plate with regularly arranged circular targets on the surface on the mobile platform of a high-precision electric displacement stage. The spatial position of the measurement plate is relatively parallel to the imaging plane of the camera. Operation The controller makes the measurement plate move a fixed distance multiple times and in a direction, and at the same time obtains a three-dimensional image of the measurement plate at each displacement position.

Moreover, the specific steps of S8 are:

S8.1: Measure the fixed-point distance length in six directions for any control point within the volume of the spatial measurement field. The error function based on the spatial distance constraint is designed as:

Lq(q=1,2,…,6) represents the measured length between the specified control point and the q-th feature point around it, and the true value of its length is Dq(q=1,2,…,6), i and j respectively represent the displacement sequence of the measurement plate and the order of the center points of the characteristic points on the measurement plate;

S8.2: Definition are the homogeneous coordinates of a set of matching points. The distances between the two points and their corresponding epipolar lines l _l and l _r are dis _l and dis _r respectively. The error function based on the epipolar constraint is constructed as:

in, for/> The corresponding left polar line,/> for/> The corresponding right pole line, F represents the basic matrix of the virtual binocular system, that is: F=R·S, R=[r ₁₁ , r ₁₂ , r ₁₃ ; r ₂₁ , r ₂₂ , r ₂₃ ; r ₃₁ , r ₃₂ , r ₃₃ ] represents the virtual binocular rotation matrix, and S represents the antisymmetric matrix of the virtual binocular translation matrix T = [t ₁ , t ₂ , t ₃ ]:

S8.3: Combine two error functions to form an objective function:

H _obj (x)＝arg min(E _L +E _P )

_{The input} variables _{of the objective} function _{are structural parameters , that is :} , combined with the SQP (Sequential Quadratic Programming) nonlinear optimization operator to iteratively solve the structural parameters, and the iteration cutoff is based on the local minimum value of the objective function output.

Moreover, the specific steps of S3 are:

S3.1: Divide the original stereo image of the calibration plate evenly along the center line of the vertical direction of the image to obtain two calibration plate sub-images I _L and I _R with the same pixel size. The left and right sub-images are regarded as left and right respectively. Captured by virtual cameras _VCL and _VCR ;

S3.2: Specify the world coordinates of feature points on the calibration plate And extract the pixel coordinates corresponding to the feature point/> Establish the mapping relationship between the 2D pixel coordinates and the 3D world coordinates of the feature points on each calibration plate image. Taking the left virtual camera as an example, the conversion relationship between the world coordinates and the pixel coordinates of the feature points is obtained:

Among them, i and j respectively represent the captured calibration plate sequence and the feature point sequence on the calibration plate, K _L is the internal parameter matrix of the left virtual camera, for/> Represents the rotation matrix and translation matrix of the optical center of VC _L relative to the i-th calibration plate;

S3.3: Use Zhang Zhengyou’s calibration method to obtain the internal parameter matrices and distortion coefficients _of VCL _{and VCR} , as well as the coordinate conversion relationship between the optical centers of _VCL and _VCR corresponding to each calibration plate image, that is: the virtual binocular structure Parameters, the average result of multiple sets of virtual binocular structural parameters is used as the system calibrated virtual binocular structural parameters.

Moreover, the structural parameters obtained in each iteration will be combined with the virtual camera internal parameters obtained during rough calibration to update the three-dimensional coordinates of the control points in the spatial measurement field. The new structural parameters and the three-dimensional coordinates of the feature point array will be used for the next time iterative calculation.

Moreover, in order to correct lens distortion, traditional radial distortion and tangential distortion are introduced into the virtual binocular vision system. Taking the left virtual camera as an example, the distortion model is defined as:

in and m _nl (x _lm ,y _lm ) are the coordinates of the distortion point and the de-distortion point on the normalized plane respectively, k _l(1-2) and p _l(1-2) respectively represent the radial and tangential distortion Coefficient,/>

An electronic device, characterized in that: it includes a processor and a memory, the memory and the processor are connected to each other, the memory stores computer instructions, and the processor executes any of the above by executing the computer instructions. method described in one item.

The advantages and positive effects of the present invention are:

The invention is suitable for calibration work of a virtual binocular vision system based on biprisms. Since the biprism virtual binocular system is a special binocular vision system, the calibration effect of traditional binocular calibration methods is relatively limited when applied to the biprism virtual binocular system. In addition, although binocular vision calibration methods based on objective function design can improve calibration accuracy, these methods also have obvious shortcomings. For example, the use of complex constraints cannot determine the type of constraints that contribute significantly to the improvement of calibration accuracy. How easy it is to cause coupled solutions. Therefore, the present invention specifically has the following advantages:

(1) The calibration method proposed by the method of the present invention fully combines the characteristics of the biprism virtual binocular vision system, performs calibration and optimization of the visual system within an artificially defined limited space, and improves the calibration quality.

(2) The objective function design based on structural parameter optimization proposed by the method of the present invention is simpler. Combining spatial distance constraints and epipolar constraints meets the comprehensive needs of the biprism virtual binocular system in terms of three-dimensional measurement and two-dimensional epipolar geometric constraints.

Description of drawings

Figure 1 is a schematic diagram of the imaging principle of the biprism virtual binocular system;

Figure 2 is an overall flow chart of the implementation of the biprism virtual dual target determination method in the embodiment of the present invention;

Figure 3 is a schematic diagram of constructing a spatial measurement field in an embodiment of the present invention;

Figure 4 is a flow chart of the nonlinear optimization method of structural parameters in the embodiment of the present invention;

Figure 5 is a schematic diagram of spatial distance error measurement in the embodiment of the present invention;

Figure 6 is a schematic diagram of the epipolar constraint error in the embodiment of the present invention. (a) is a schematic diagram of epipolar geometry, and (b) is a diagram of epipolar distance definition.

Detailed ways

In order to make the purpose, technical solutions and advantages of the present application clearer and easier to understand, the present invention will be further described below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described here are only used to explain the application and are not used to limit the application.

The present invention provides a structural parameter optimization calibration method for a biprism virtual binocular vision system. The overall calibration process is shown in Figure 2. The overall process of the present invention is as follows:

Step 1: Build a virtual binocular stereoscopic vision system based on biprism refraction. The relative position of the camera and the biprism is adjusted through a precision shift stage so that the biprism is relatively parallel to the imaging surface of the camera and the visual system obtains evenly divided left and right sub-images. The imaging principle of the biprism virtual binocular vision system is shown in Figure 1. The acquired stereoscopic image is evenly divided along the center line of the image to obtain two sub-images with the same pixel size. The left and right sub-images can be regarded as taken by two virtual cameras on the left (VC _L ) and right (VC _R ) at different positions.

Step 2: Randomly place the calibration plate with solid dots on the surface at different positions in the common field of view of the virtual binoculars to obtain the stereoscopic calibration plate image.

Step 3: Evenly divide the calibration plate image obtained in Step 2 along the center line of the image, and extract the three-dimensional coordinates of the center feature points on each calibration plate. and two-dimensional pixel coordinates/> Taking the left virtual camera as an example, establish the mapping relationship between the world coordinates and pixel coordinates of the calibration plate feature points:

Among them, i and j respectively represent the photographed calibration plate sequence and the feature point sequence on the calibration plate. K _L is the internal parameter matrix of the left virtual camera, for/> Represents the rotation matrix and translation matrix of the optical center of VC _L relative to the i-th calibration plate. In addition, in order to correct lens distortion, this example introduces traditional radial distortion and tangential distortion into the virtual binocular vision system. The distortion model is:

in is the coordinate of the distortion point on the normalized plane, m _nl (x _lm , y _lm ) is the coordinate of the dedistorted point on the normalized plane. k _l(1-2) and p _l(1-2) respectively represent the radial and tangential distortion coefficients,/>

Step 4: Based on the virtual camera imaging model described in Step 3, use Zhang Zhengyou’s calibration method to roughly calibrate the left and right virtual cameras. Coarse calibration can obtain the internal parameter matrices, distortion coefficients of each of the two virtual cameras, and the coordinate transformation relationship between the VC _L and VC _R optical centers corresponding to each calibration plate image, that is, the virtual binocular structural parameters. In the present invention, the average value of multiple sets of structural parameters is used as the initial value of the structural parameters for system calibration.

Step 5: As shown in Figure 3, a measuring plate with a surface number of 10×21 and a regular arrangement is vertically fixed on a high-precision displacement stage. The displacement direction of the measuring plate is relatively perpendicular to the imaging plane of the camera. The displacement stage is controlled to move the measurement plate 1mm each time for a total of 51 times. Each circle center feature point on the measurement plate is regarded as a control point. According to the linear displacement of the measurement plate, a spatial measurement field with a standardized arrangement and a known distance between adjacent control points is constructed in the common field of view of the virtual binoculars.

Step 6: Uniformly segment each group of stereo images used to construct the spatial measurement field. According to the internal reference results obtained in step 4, perform distortion correction on the left and right view measurement plate images and extract the pixel coordinates of the control points. The matching coordinates of the control points in the spatial control field are obtained according to the arrangement order of the control points on the measurement plate and the displacement order of the measurement plate in the three-dimensional space. Then the objective function is constructed to perform nonlinear optimization of the structural parameters. The specific process is shown in Figure 4.

Step 7: Use the linear trigonometric method to recover the three-dimensional coordinates of each control point with the optical center of the left camera as the coordinate system, thereby establishing a constraint function based on the spatial distance error. The definition of spatial distance error is shown in Figure 5. Select a certain control point in the measurement space, and measure the control point and the point M1 in the left diagonal upward direction, the point M2 in the right diagonal upward direction, and the left diagonal downward direction on the plane of the measurement plate. The spatial distances between point M3, point M4 in the right and downward direction, and six line segments relative to the forward point M5 and the backward point M6 of the measurement board where they are located. The error function based on spatial distance constraints is designed as:

Lq (q=1,2,...,6) represents the measured length between the specified control point and the qth control point around it, and the true value of its length is Dq (q=1,2,...,6). i and j respectively represent the displacement sequence of the measurement plate and the sequence of control points on the measurement plate.

Step 7: Design the constraint function based on epipolar constraints. Figure 6(a) shows the epipolar geometric model of the binocular vision system. The optical centers of _VCL and _VCR are _Ocl and _Ocr . The line segment connecting _Ocl and _Ocr is defined as the baseline of the binocular system. Its intersection points with the imaging surface of the left virtual camera and the imaging surface of the right virtual camera are respectively , right pole: e _l , e _r . is the homogeneous coordinates of the matching point pair on the pixel coordinate system. The left polar line and right polar line corresponding to the matching point can be expressed as:/> F represents the basic matrix of the virtual binocular system, that is: F=R·S.

R＝[r ₁₁ , r ₁₂ , r ₁₃ ; r ₂₁ , r ₂₂ , r ₂₃ ; r ₃₁ , r ₃₂ , r ₃₃ ] represents the virtual binocular rotation matrix, S represents the translation matrix T＝[t ₁ , t ₂ ,t ₃ ] antisymmetric matrix:

Ideally, the matching point of the specified point should be on the epipolar line calculated from the pixel coordinates of the specified point, that is, the distance from the matching point to its corresponding epipolar line is 0. But this is not the actual situation. There is always a certain vertical distance between the matching point and its corresponding epipolar line. The smaller the vertical distance, the more the structural parameters are in line with the epipolar geometric constraints of the current stereo vision system. Figure 6(b) shows the schematic diagram of the epipolar distance error, defined The distances to the polar lines l _l and l _r are dis _l and dis _r respectively. The error function based on epipolar constraints is designed as:

Among them, i and j respectively represent the displacement sequence of the measurement plate and the arrangement order of the control point array on the measurement plate.

Step 8: The above steps 6 and 7 construct an error constraint function based on spatial distance error and epipolar constraints. Combine the two constraint functions and construct the objective function as:

H _obj (x) = arg min ( _EL + E _P ).

_{Here , there are 12} variables _{that need to} be optimized _{, namely : t3} ). The objective function is solved using the nonlinear optimization operator Sequential Quadratic Programming (SQP) method. The nonlinear optimization of structural parameters is an iterative process, and the error output obtained from each iteration will be used to judge the iteration cutoff conditions. If the iteration cutoff conditions are not met, the structural parameters output from this iteration will be used as the initial values for the next iteration, and the three-dimensional coordinates of the control points in the spatial measurement field will be updated based on the current structural parameters. When the objective function output reaches the cut-off condition, the current structural parameters are considered to have reached the local optimal value of nonlinear optimization.

The invention also provides an electronic device, including:

A memory and a processor. The memory and the processor are communicatively connected to each other. Computer instructions are stored in the memory, and the processor executes any of the above methods by executing the computer instructions.

The present invention also provides a computer storage medium. Computer instructions are stored on the computer-readable storage medium, and the computer instructions are used to cause the computer to execute any of the above methods.

Those skilled in the art will understand that 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 that combines 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 solutions in the embodiments of this application can be implemented using various computer languages, such as the object-oriented programming language Java and the literal scripting language JavaScript.

The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the application. It will be understood that each process and/or block in the flowchart illustrations and/or block diagrams, and combinations of processes and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing device to produce a machine, such that the instructions executed by the processor of the computer or other programmable data processing device produce a use A device for realizing the functions specified in one process or multiple processes of the flowchart and/or one block or multiple blocks of the block diagram.

These computer program instructions may also be stored in a computer-readable memory that causes a computer or other programmable data processing apparatus to operate in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including the instruction means, the instructions The device implements the functions specified in a process or processes of the flowchart and/or a block or blocks of the block diagram.

These computer program instructions may also be loaded onto a computer or other programmable data processing device, causing a series of operating steps to be performed on the computer or other programmable device to produce computer-implemented processing, thereby executing on the computer or other programmable device. Instructions provide steps for implementing the functions specified in a process or processes of a flowchart diagram and/or a block or blocks of a block diagram.

The above-mentioned embodiments only express several implementation modes of the present application. The descriptions are relatively specific and detailed, but should not be construed as limiting the patent scope of the present invention. It should be noted that, for those of ordinary skill in the art, several modifications and improvements can be made without departing from the concept of the present application, and these all fall within the protection scope of the present application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims (10)

1. A structural parameter optimization calibration method for a biprism virtual binocular vision system is characterized by comprising the following steps of:

s1: adjusting the relative position of the camera and the biprism to enable the biprism virtual binocular vision system to acquire left and right sub-images which are evenly divided;

s2: the calibration plates are placed at different positions under the depth of field range of the biprism virtual binocular vision system, and three-dimensional images of a plurality of groups of calibration plates are obtained;

s3: uniformly dividing an original stereoscopic image of a calibration plate to obtain two sub-images with the same size, respectively extracting characteristic points on a left sub-image and a right sub-image, roughly calibrating a biprism virtual binocular vision system, and obtaining an internal reference and a structural parameter initial value of a virtual binocular camera;

s4: the measuring plate is placed in front of the vision system, the measuring plate is moved for a plurality of times at a fixed distance, a stereoscopic image sequence is obtained, and a space measuring field composed of characteristic points with regular distribution is established;

s5: uniformly dividing a plurality of groups of stereoscopic images in a space measurement field, and carrying out distortion correction on sub-images corresponding to the left and right virtual cameras by utilizing the internal parameters of the virtual binocular cameras obtained in the step S3;

s6: extracting circle center characteristic points on the distorted measurement plate image, and matching the circle center characteristic points of the left and right subgraphs according to the arrangement sequence of the circle center characteristic point sequences to obtain characteristic point matching point pairs;

s7: combining internal parameters and structural parameters of the virtual binocular camera, and recovering three-dimensional coordinates of characteristic points in the space measurement field by a linear trigonometry;

s8: and constructing an objective function based on the space distance error and epipolar constraint, performing nonlinear iterative optimization on the structural parameters, and updating the three-dimensional coordinates of the feature points in the step S7 by using the structural parameters obtained by each iterative calculation.

2. The structural parameter optimization calibration method for a biprism virtual binocular vision system of claim, wherein the method comprises the following steps: the biprism virtual binocular vision system is a binocular stereoscopic vision image pair which is obtained by a single camera after being refracted by a biprism and can be equivalent to binocular stereoscopic vision images shot by two virtual cameras.

3. The structural parameter optimization calibration method for a biprism virtual binocular vision system according to claim 1, wherein: the specific steps of the S1 are as follows: the positions of the camera and the biprism are adjusted by the precise displacement table, so that the center line in the vertical direction of the image is approximately overlapped with the edge of the center of the biprism.

4. The structural parameter optimization calibration method for a biprism virtual binocular vision system according to claim 1, wherein: and (5) performing coarse calibration on the biprism virtual binocular vision system by using a Zhang Zhengyou calibration method.

5. The structural parameter optimization calibration method for a biprism virtual binocular vision system according to claim 1, wherein: the specific steps of the S4 are as follows: a measuring plate with a circular target arranged regularly on the surface is fixed on a moving platform of a high-precision electric displacement table, the spatial position of the measuring plate is relatively parallel to an imaging surface of a camera, and an operation controller enables the measuring plate to move a fixed distance for a plurality of times and directionally, and meanwhile, a three-dimensional image of the measuring plate at each displacement position is acquired.

6. The structural parameter optimization calibration method for a biprism virtual binocular vision system according to claim 1, wherein: the specific steps of the S8 are as follows:

s8.1: the fixed point distance length measurement in six directions is carried out on any one control point in the volume of the space measurement field, and an error function based on space distance constraint is designed as follows:

lq (q=1, 2, … …, 6) represents the measured length between the designated control point and the q-th feature point around it, the true value of the length is Dq (q=1, 2, … …, 6), i and j represent the displacement order of the measuring plate and the center of circle order of the feature point on the measuring plate, respectively;

s8.2: definition of the definitionIs the homogeneous coordinates of a group of matching points, and the two points are connected to the corresponding polar line l _l And/l _r Is dis respectively _l 、dis _r Constructing an error function based on epipolar constraint as follows:

wherein,is->Corresponding left pole line,/->Is->The corresponding right epipolar line, F, represents the basis matrix of the virtual binocular system, namely: f=r·s, r= [ R ] ₁₁ ,r ₁₂ ,r ₁₃ ；r ₂₁ ,r ₂₂ ,r ₂₃ ；r ₃₁ ,r ₃₂ ,r ₃₃ ]Representing a virtual binocular rotation matrix, S representing a virtual binocular translation matrix t= [ T ] ₁ ,t ₂ ,t ₃ ]Is an antisymmetric matrix of (a):

s8.3: combining two error functions to form an objective function:

H _obj (x)＝arg min(E _L +E _P )

the input variables of the objective function are structural parameters, namely: x= (r ₁₁ ,r ₁₂ ,r ₁₃ ,r ₂₁ ,r ₂₂ ,r ₂₃ ,r ₃₁ ,r ₃₂ ,r ₃₃ ,t ₁ ,t ₂ ,t ₃ ) And carrying out iterative solution on the structural parameters by combining SQP (Sequential Quadratic Programming) nonlinear optimization operators, wherein the iterative cutoff is based on the output local minimum value of the objective function.

7. The structural parameter optimization calibration method for the biprism virtual binocular vision system according to claim 4, wherein the structural parameter optimization calibration method is characterized by comprising the following steps of: the specific steps of the S3 are as follows:

s3.1: uniformly dividing an original stereoscopic image of the calibration plate along a central line in the vertical direction of the image to obtain two calibration plate sub-images I with the same pixel size _L 、I _R The left and right sub-images are respectively regarded as left and right virtual cameras VC _L 、VC _R Shooting the obtained product;

s3.2: world coordinates defining feature points on calibration plateAnd extracting the pixel coordinates corresponding to the feature points Establishing a mapping relation between 2D pixel coordinates and 3D world coordinates of the feature points on each calibration plate image, and taking a left virtual camera as an example, obtaining a conversion relation between the feature point world coordinates and the pixel coordinates:

wherein i and j respectively represent the shot calibration plate sequence and the characteristic point sequence on the calibration plate, K _L Is an internal reference matrix of the left virtual camera,is->Representing VC _L A rotation matrix and a translation matrix of the optical center of the (c) relative to the ith calibration plate;

s3.3: VC is obtained by adopting Zhang Zhengyou calibration method _L ，VC _R Is characterized by comprising an internal reference matrix, distortion coefficients and VC corresponding to each calibration plate image _L With VC _R Coordinate conversion relation between optical centers, namely: and the virtual double-purpose structure parameters are used as virtual double-purpose structure parameters for system calibration by taking the average value of the multiple groups of virtual double-purpose structure parameters.

8. The structural parameter optimization calibration method for the biprism virtual binocular vision system according to claim 6, wherein: the structural parameters obtained by each iteration are combined with the virtual camera internal parameters obtained during coarse calibration to update the three-dimensional coordinates of the control points in the space measurement field, and the new structural parameters and the three-dimensional coordinates of the feature point array are used for the next iteration calculation.

9. The structural parameter optimization calibration method for a biprism virtual binocular vision system according to claim 7, wherein: to correct lens distortion, conventional radial distortion and tangential distortion are introduced into a virtual binocular vision system, taking a left virtual camera as an example, the distortion model is defined as:

wherein the method comprises the steps ofAnd m is equal to _nl (x _lm ,y _lm ) The coordinates of the distortion point and the de-distortion point on the normalized plane are respectively k _l(1-2) And p is as follows _l(1-2) Each representing a radial and tangential distortion coefficient, +.>

10. An electronic device, characterized in that: comprising a processor and a memory, said processors being communicatively coupled to each other, said memory having stored therein computer instructions, said processor executing the method of any of claims 1 to 9 by executing said computer instructions.