Cylinder mark point eccentricity correction algorithm space pose tracking method and device

Abstract

The invention discloses a space pose tracking method and device based on a cylinder mark point eccentricity correction algorithm, and relates to the technical field of cylindrical part pose tracking, and the method comprises the steps: firstly, carrying out the calibration of a right-angle stereo target image through a binocular vision measurement device, obtaining parameter information, and carrying out the image collection of a cylindrical part; then three-dimensional reconstruction and discrete sampling are carried out according to the collected image, and an expanded view after geometric distortion correction is generated; and then positioning the center coordinate of the mark point through an interpolation algorithm, solving the three-dimensional coordinate of the coordinate through binocular triangulation and linear transformation, and finally outputting the spatial pose parameter of the measured cylindrical part. According to the invention, the limitation of the prior art on the measurement of the solid cylindrical part is broken through, the circular mark points are arranged on the outer surface of the cylindrical part, an eccentricity correction algorithm is provided, the problem of mark point positioning deviation caused by curved surface distortion is solved, the pose measurement precision is improved, and reliable technical support is provided for the precise assembly of the cylindrical part.

Description

Technical Field

[0001] The present invention relates to the technical field of cylindrical part posture tracking, and in particular to a method and device for spatial posture tracking of a cylindrical marker point eccentricity correction algorithm. Background Art

[0002] Cylindrical structural components are widely used in machinery manufacturing, aerospace, energy equipment, and other fields. High-precision tracking of their spatial pose is a core requirement for automated assembly, quality inspection, and motion control. While current technologies have achieved a certain level of measurement accuracy, their versatility is still limited. Measuring devices that rely on internal operations are not suitable for solid cylindrical components such as engine crankshafts, while pre-installed tooling solutions are complex to deploy and lack flexibility. Furthermore, surface distortion increases errors in collected data, which in turn affects the accuracy of pose calculations.

[0003] In the prior art, patent CN202210182068.0 proposes a pipeline posture measurement device, which includes a guide power assembly, three variable diameter bracket assemblies, a circumferential rotation assembly, a detection assembly, and a controller. The method is as follows: after the laser displacement sensor rotates circumferentially at any angle driven by the circumferential rotation assembly, a number of laser points are projected on the inner wall of the pipeline. In the algorithm, these laser points are subjected to coordinate transformation and ellipse fitting to determine the axis vector of the ellipse. Finally, the relative posture relationship between the two pipeline axes is calculated in the controller to achieve posture estimation of large pipelines. Although this method can achieve pipeline posture measurement, it strictly relies on internal measurement devices and cannot be applied to solid cylindrical parts.

[0004] Patent CN202311794877 proposes a method for automatically grasping and flexibly docking large-mass cylinders, including the following steps: S1-S5: Installing tooling and placing the cylindrical assembly; S6-S9: A visual camera identifies feature points on the tooling, calculates the pose, and binds them to the robot end; S10-S14: The robot grasps the product, aligns the axis with the pin hole, and fine-tunes the force control system to achieve a smooth docking. While the introduction of vision and force control improves docking success rates and avoids product damage, the need for pre-installed specialized tooling complicates deployment and prevents direct application to cylinders without tooling. The visual system relies on tooling landmarks, and the unprocessed cylindrical surface causes imaging distortion at these landmarks, reducing pose calculation accuracy. Summary of the Invention

[0005] The first purpose of the present invention is to provide a spatial posture tracking method with a cylindrical marker eccentricity correction algorithm, aiming to solve the problem that the existing methods lack a universal measurement mechanism for solid / hollow cylindrical parts, and the surface distortion leads to increased errors in the collected data, which in turn affects the accuracy of the posture solution.

[0006] In order to solve the above technical problems, a method for spatial posture tracking using a cylindrical marker eccentricity correction algorithm is provided, comprising the following steps:

[0007] S1. Calibrate the rectangular stereo target image using a binocular vision measurement device to obtain parameter information, and collect images of the cylindrical part;

[0008] S2, performing three-dimensional reconstruction and discrete sampling based on the acquired image to generate an expanded image after geometric distortion correction;

[0009] S3, locate the center coordinates of the marker point through interpolation algorithm;

[0010] S4, solving the three-dimensional coordinates of the center coordinates of the marker point through binocular triangulation and linear transformation;

[0011] S5. Output the spatial posture parameters of the cylindrical part being measured.

[0012] In one embodiment, a binocular vision measurement device is built, and a rectangular stereo target image is placed to calibrate the binocular vision measurement device, and the internal parameters and external parameters of the left camera and the right camera and the pose relationship between the two cameras are obtained, wherein the internal parameters include the left camera intrinsic parameter matrix K l , the right camera intrinsic parameter matrix K r , the left camera distortion coefficient k li and the right camera distortion coefficient k ri (i=1,2), the external parameters include the rotation matrix R of the left camera relative to the world coordinate system wl With the translation vector t wl , the rotation matrix R of the right camera relative to the world coordinate system wr With the translation vector t wr , the pose relationship between the left camera and the right camera includes the rotation matrix R lr and the translation vector t lr , paste circular marking points on the surface of the cylindrical part, collect the side image of the cylindrical part through the binocular vision measurement device, and apply the left camera distortion coefficient k li and the right camera distortion coefficient k ri The collected side image of the cylinder is dedistorted.

[0013] In one embodiment, the radius r of the cylindrical part is obtained, the EDLines line detection algorithm is used to extract the cylindrical part contour lines l1 and l2 in the side image of the cylindrical body, and the back-projection plane normal vector n corresponding to the contour lines is calculated. i :

[0014]

[0015] Solve for the bisecting plane normal vector n3:

[0016]

[0017] Among them, ||n i || represents the back-projected plane normal vector n i The modulus of the geometric parameters is then calculated, and the angle θ of the back-projection plane corresponding to the contour line, the direction vector s of the cylinder axis L , and the optical center O of the binocular vision measurement device C To the center P C The direction vector s OP :

[0018]

[0019] Determine the optical center O of the binocular vision measurement device C To the center P C The distance is:

[0020]

[0021] Calculate the center P of the circle C Coordinates:

[0022]

[0023] The three-dimensional model of the cylindrical part is completed.

[0024] In one embodiment, in the binocular vision measurement device coordinate system X C Y C Establish a standard circle C on the plane and discretize the central angle δ to obtain a point set Arc length The discretization process is performed along the axis of the cylinder to obtain the cross-section circle sequence {C j |j=1,2,…,J}, with the center of the cross-section circle P cj As the origin, the cylinder axis direction vector s L is the X axis, the bisecting plane normal vector n3 is the Y axis, and the direction vector s OP For the Z axis, establish a cross-section circle coordinate system, and use the transformation matrix between the binocular vision measurement device coordinate system and the cross-section circle coordinate system:

[0025]

[0026] Calculation of cross-section circle C j Discrete points on Coordinates on the standard circle C

[0027]

[0028] in Respectively and The homogeneous coordinates of .

[0029] In one embodiment, the discrete points are calculated The optical center O of the binocular vision measurement device C The vector angle of :

[0030]

[0031] Where T1 is the contour tangent point. When β < 90°, it is determined to be a visible discrete point. The visible discrete point passes through:

[0032]

[0033] Get the coordinates of its projection point on the image in for The homogeneous coordinates, I is the identity matrix, and the cross-section circle C j The discrete point index i on the image is the row coordinate of the expanded image, and the index j is the column coordinate of the expanded image. Corresponding grayscale information Mapping to the corresponding position in the unfolded view matrix to generate an unfolded view of the outer surface of the cylindrical part.

[0034] In one embodiment, the EDCircles algorithm is used to detect the circular mark in the flattened image and extract the initial circle center coordinates p fc =[u fc v fc ] T , locate the initial circle center coordinate p fc The four vertices of the pixel grid:

[0035]

[0036] Extract the decimal part of the coordinates:

[0037]

[0038] Compute exact circular coordinates via bilinear interpolation:

[0039]

[0040] What you want (u c ,v c ) are the center coordinates of the circular marking point on the outer surface of the cylindrical member.

[0041] In one embodiment, the optical ray O emitted by the left camera is calculated and tracked through the above steps.l P l and the optical ray O emitted by the right camera r P r Intersection point P w , based on the linear transformation algorithm, the least squares estimation of the three-dimensional coordinates of the spatial point is completed:

[0042]

[0043] Among them, B is the coefficient matrix of the equation system, (u l ,v l ) and (u r ,v r ) are the intersection points P w Project point p on the left and right images l and p r The image coordinates of and is the projection matrix M l =K l [R wl t wl ] and M r =K r [R wr t wr ], solve the system of equations by linear least squares method Find P w An approximate solution to obtain the intersection point P w The three-dimensional coordinates in the camera coordinate system are the position and posture of the cylinder in the camera coordinate system.

[0044] The second object of the present invention is to provide a spatial posture tracking device with a cylindrical marker eccentricity correction algorithm, aiming to solve the problems that the existing devices do not construct a collaborative processing architecture of a binocular vision measurement device and an eccentricity correction algorithm, and cannot achieve high-precision posture solution for all types of cylindrical parts while simplifying the hardware structure; lack a dynamic compensation module for marker point imaging distortion, and cannot eliminate positioning deviations caused by curvature and viewing angle; and do not integrate a lightweight data processing engine, and cannot achieve efficient real-time measurement in an industrial field environment.

[0045] In order to solve the above technical problems, a cylindrical marker point eccentricity correction algorithm spatial posture tracking device is provided, which runs the above cylindrical marker point eccentricity correction algorithm spatial posture tracking method, including:

[0046] Image processing module, including binocular vision measurement device, tripod, operation processor, the binocular vision measurement device includes left camera and right camera, the tripod is connected to the bottom of the binocular vision measurement device, the binocular vision measurement device is used to collect the image of the cylindrical piece to be measured, and the operation processor is arranged in the binocular vision measurement device for processing output required parameters;

[0047] Marking point attaching module, including sliding platform, cylindrical piece, circular marking point, a plurality of circular marking points are distributed on the surface of the cylindrical piece, the cylindrical piece is laid on the top of the sliding platform, and the relative position of the cylindrical piece and the binocular vision measurement device is adjusted through the sliding platform.

[0048] The embodiment of the application has the following beneficial effects:

[0049] 1. The cylindrical marking point eccentricity correction algorithm space pose tracking method in the embodiment one, first, the binocular vision measurement device is used to collect the cylindrical piece side image, and the double target positioning is completed based on the right-angle stereo target, the camera internal and external parameters and the pose relationship are obtained, then the contour line is extracted by using the cylindrical contour reconstruction algorithm, the axis direction and the center position are calculated by combining the back projection plane normal vector, and the cylindrical three-dimensional model is constructed, then the geometric distortion correction development diagram is generated by discrete sampling, and the imaging distortion of the marking point caused by the curvature is eliminated, then the sub-pixel level positioning of the marking point center is realized by using the bilinear interpolation, finally, the three-dimensional coordinates of the marking point are solved by combining the binocular triangulation and the straight line linear transformation algorithm, and the high-precision space pose parameters are output. Compared with the prior art, the present scheme combines the three core technical advantages of the outer surface marking point eccentricity correction algorithm, the binocular vision general measurement architecture and the lightweight data processing engine, and innovatively introduces the dynamic distortion compensation mechanism based on the development diagram, and systematically solves the four technical bottlenecks of the traditional method, that is, the limitation of solid piece measurement, the complex device structure, the poor environmental robustness and the low calculation efficiency. Therefore, the pose tracking accuracy and stability of the cylindrical piece in the complex industrial environment are significantly improved, the invasive measurement and the hardware dependence of multi-sensor cooperation are avoided, and the general, efficient and robust measurement of solid / hollow cylindrical pieces is realized.

[0050] 2. The cylindrical marker point eccentricity correction algorithm space pose tracking device in the second embodiment, through the cooperative work of the image acquisition and processing module and the measurement module, a hardware-software cooperative architecture integrating image acquisition-distortion correction-pose solution is constructed. Among them: the image acquisition and processing module: the marker point image is synchronously acquired through the binocular vision measurement device, and the geometric distortion elimination and marker point accurate positioning are realized by combining the eccentricity correction algorithm; the measurement module: the cylindrical part is fixed through the sliding platform to ensure the imaging stability of the marker point; the operation processor: the discrete sampling visualization judgment, sub-pixel interpolation positioning and binocular triangulation algorithm are integrated, and the pose parameters are output in real time. Compared with the existing device, the scheme realizes the tight coupling design of the binocular vision measurement device and the dynamic distortion compensation, especially the integration of the eccentricity correction algorithm of the outer surface marker point and the lightweight processing engine, which effectively solves the problems of solid part measurement failure, complex environment precision error and real-time deficiency caused by the lack of curvature adaptive compensation mechanism in the existing device. The device can realize high-precision pose tracking of the cylindrical part under the environment of changing light, surface reflection or shielding, and simplify the hardware structure, providing a reliable and efficient solution for industrial automation assembly. BRIEF DESCRIPTION OF DRAWINGS

[0051] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed to be used in the embodiment or prior art description will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0052] Figure 1 The flowchart of the cylindrical marker point eccentricity correction algorithm space pose tracking method according to the first embodiment of the present application;

[0053] Figure 2 The structure diagram of the cylindrical marker point eccentricity correction algorithm space pose tracking device according to the second embodiment of the present application;

[0054] Figure 3 The principle diagram of the binocular geometric measurement model according to the first embodiment of the present application.

[0055] 1, cylindrical part; 2, circular marker point; 3, binocular measurement device; 4, tripod; 5, sliding platform. DETAILED DESCRIPTION

[0056] Hereinafter, only certain exemplary embodiments are briefly described. As those skilled in the art will appreciate, the described embodiments may be modified in various ways without departing from the spirit or scope of the embodiments of the present invention. Therefore, the drawings and descriptions are considered to be illustrative and non-restrictive in nature. The disclosure below provides many different embodiments or examples for implementing different structures of the embodiments of the present invention. In order to simplify the disclosure of the embodiments of the present invention, the components and settings of specific examples are described below. Of course, they are merely examples and are not intended to limit the embodiments of the present invention. In addition, the embodiments of the present invention may repeat reference numerals and / or reference letters in different examples, and such repetition is for the purpose of simplicity and clarity and does not in itself indicate the relationship between the various embodiments and / or settings discussed. The embodiments of the present invention are described in detail below in conjunction with the accompanying drawings.

[0057] Example 1

[0058] Please refer to Figure 1 、 2 A first embodiment of the present invention provides a spatial pose tracking method using a cylindrical marker eccentricity correction algorithm. This embodiment includes the steps of setting up a binocular vision measurement device. First, the binocular measurement device 3 is fixed to a tripod 4. The camera baseline distance and focal length are adjusted to ensure that the field of view covers the cylindrical member 1 on the sliding platform 5. High-contrast circular markers 2 are evenly attached to the outer surface of the cylindrical member 1. The sliding platform 5 is moved to three different spatial poses, and the binocular measurement device 3 is synchronously triggered to collect data.

[0059] Binocular vision measurement device calibration steps, based on the rectangular stereo target image, calculate the left and right camera intrinsic parameter matrix K l ,K r , distortion coefficient d l ,d r , as well as the pose relationship between the two cameras, and apply the distortion correction model to process the collected cylindrical images.

[0060] The steps of cylinder contour reconstruction and discrete sampling to generate the unfolded image for geometric distortion correction are as follows: the radius r of the cylinder is obtained, the EDLines line detection algorithm is used to extract the cylinder contour lines l1 and l2 in the cylinder side image, and the back-projection plane normal vector n corresponding to the contour lines is calculated. i :

[0061]

[0062] Solve for the bisecting plane normal vector n3:

[0063]

[0064] Calculate the angle θ of the back-projection plane corresponding to the geometric parameter contour line and the cylinder axis direction vector sL , and the optical center O of the binocular vision measurement device C To the center P C The direction vector s OP :

[0065]

[0066] Determine the optical center O of the binocular vision measurement device C To the center P C The distance is:

[0067]

[0068] Calculate the center P C Coordinates:

[0069]

[0070] Complete the 3D model of the cylindrical part.

[0071] In the camera coordinate system X C Y C Establish a standard circle C on the plane and discretize the central angle δ to obtain a point set Arc length along the axis Discretize and generate a cross-section circle sequence {C j |j=1,2,…,J}; establish the cross-section circle coordinate system (origin: cross-section circle center P cj , X axis: cylinder axis direction vector s L , Y axis: bisecting plane normal vector n3, Z axis: direction vector s OP ), through coordinate transformation:

[0072]

[0073] Calculation of cross-section circle C j Discrete points on Coordinates on the standard circle C Then make a visual judgment and calculate the discrete points Optical center O of binocular vision measurement device C The vector angle of :

[0074]

[0075] When β < 90°, it is determined to be a visible discrete point, and the visible discrete point passes through:

[0076]

[0077] Get the coordinates of its projection point on the image And the cross section circle C jThe discrete point index i on the image is the row coordinate of the expanded image, and the index j is the column coordinate of the expanded image. Corresponding grayscale information Map to the corresponding position in the unfolded view matrix to generate the unfolded view of the outer surface of the cylindrical part.

[0078] In the step of locating the center coordinates of the marker points, the EDCircles algorithm is used to detect the circular marker points in the flattened image and extract the initial circle center coordinates p fc =[u fc v fc ] T , locate pixel grid vertices:

[0079]

[0080] Bilinear interpolation is used to calculate the center coordinates of the circular marking points on the outer surface of the cylinder:

[0081]

[0082] Three-dimensional coordinate solution steps, through the above steps to calculate and track the optical ray O emitted by the left camera l P l and the optical ray O emitted by the right camera r P r Intersection point P w , construct the system of equations:

[0083]

[0084] Solve the system of equations by linear least squares method Find P w An approximate solution to obtain the intersection point P w The three-dimensional coordinates in the camera coordinate system are the position and posture of the cylinder in the camera coordinate system.

[0085] Implementing the first embodiment of the present invention will have the following beneficial effects: This embodiment uses a binocular vision measurement device 3 and a sliding platform 5 to collaboratively capture multi-angle images of a cylindrical object 1; utilizes an eccentricity correction algorithm to eliminate geometric distortion and achieve sub-pixel positioning of the marker 2; combines discrete sampling with visual judgment to reduce computational complexity; and ultimately outputs high-precision pose parameters through binocular triangulation. Compared to existing technologies, this solution innovatively introduces a dynamic distortion compensation mechanism based on an unfolded graph, addressing three major issues: limited measurement of solid objects, poor environmental robustness, and low computational efficiency. This significantly improves the pose tracking accuracy and stability of cylindrical objects in complex industrial environments, while avoiding the hardware dependency of invasive measurement and multi-sensor collaboration, achieving universal, efficient, and robust measurement of solid and hollow cylindrical objects.

[0086] Example 2

[0087] The subject matter protected by the cylindrical marker eccentricity correction algorithm spatial pose tracking device in the second embodiment is different from the cylindrical marker eccentricity correction algorithm spatial pose tracking method in the first embodiment. The specific differences are as follows:

[0088] Please refer to Figure 2 , is a method for spatial posture tracking using a cylindrical marker eccentricity correction algorithm according to an embodiment of the present invention. The device includes the following modules and steps:

[0089] The image processing module includes a binocular measurement device 3, a tripod 4, and an operation processor. The binocular vision measurement device 3 includes a left camera and a right camera. The tripod is connected to the bottom of the binocular vision measurement device 3. The binocular vision measurement device is used to collect images of the cylindrical part to be measured. The operation processor is set in the binocular vision measurement device 3 to process and output the required parameters; the marking point placement module includes a sliding platform 5, a cylindrical part 1, and circular marking points 2. Multiple circular marking points 2 are distributed on the surface of the cylindrical part. The cylindrical part 1 lies on its side above the sliding platform 5. The relative position of the cylindrical part and the binocular vision measurement device 3 is adjusted by the sliding platform 5.

[0090] The module runs the following steps: S1, calibrating the rectangular stereo target image to obtain parameter information through a binocular vision measurement device, and collecting images of the cylindrical part;

[0091] S2, performing three-dimensional reconstruction and discrete sampling based on the acquired image to generate an expanded image after geometric distortion correction;

[0092] S3, locate the center coordinates of the marker point through interpolation algorithm;

[0093] S4, solving the three-dimensional coordinates of the center coordinates of the marker point through binocular triangulation and linear transformation;

[0094] S5. Output the spatial posture parameters of the cylindrical part being measured.

[0095] Through the collaborative work of the image acquisition and processing module and the measurement module, a hardware-software collaborative architecture integrating image acquisition, three-dimensional reconstruction, distortion compensation, and pose solution has been constructed. Specifically, the image acquisition module uses a sliding platform 5 to achieve multi-pose control of the cylindrical part 1, ensuring full coverage of the marker points 2; the pose solution output module combines a lightweight least squares algorithm to achieve accurate pose parameter output. Compared with existing devices, this solution effectively solves the problems of solid part measurement failure, inaccurate accuracy in complex environments, and insufficient real-time performance caused by the lack of a curvature adaptive compensation mechanism in existing devices through a tightly coupled design of a binocular vision measurement device and dynamic distortion compensation, specifically the integration of an algorithm for correcting eccentricity of external surface marker points with a lightweight processing engine. This device can achieve high-precision pose tracking of cylindrical parts in environments with varying lighting, surface reflections, or occlusion using a simplified hardware structure, providing a reliable and efficient solution for industrial automated assembly.

[0096] In some embodiments, the line structured light sensor light surface calibration device and method also include: a processor, which is respectively connected to the system assembly and image acquisition module, the light strip feature extraction and camera calibration module, the three-dimensional coordinate solution module, the virtual camera pose estimation module, and the structure parameter calibration and light surface generation module; a memory, which is connected to the processor and stores a computer program that can be run on the processor; wherein, when the processor executes the computer program, the processor controls the system assembly and image acquisition module, the light strip feature extraction and camera calibration module, the three-dimensional coordinate solution module, the virtual camera pose estimation module, and the structure parameter calibration and light surface generation module to implement any of the above-mentioned line structured light sensor light surface calibration methods.

[0097] The above embodiments are intended to illustrate the present invention, not to limit the present invention. Therefore, changes in illustrative values ​​or substitutions of equivalent components should still fall within the scope of the present invention.

Claims

1. A method for spatial posture tracking using a cylindrical marker eccentricity correction algorithm, characterized in that: Including steps: S1. Calibrate the rectangular stereo target image using a binocular vision measurement device to obtain parameter information, and collect images of the cylindrical part; S2, performing three-dimensional reconstruction and discrete sampling based on the acquired image to generate an expanded image after geometric distortion correction; S3, locate the center coordinates of the marker point through interpolation algorithm; S4, solving the three-dimensional coordinates of the center coordinates of the marker point through binocular triangulation and linear transformation; S5. Output the spatial posture parameters of the cylindrical part being measured.

2. The method for spatial posture tracking using a cylindrical marker eccentricity correction algorithm according to claim 1, characterized in that: The method of calibrating the rectangular stereo target image by a binocular vision measurement device to obtain parameter information and collecting the image of the cylindrical part includes the following steps: Build a binocular vision measurement device, place a rectangular stereo target image to calibrate the binocular vision measurement device, and obtain the internal parameters, external parameters and the pose relationship between the left and right cameras, where the internal parameters include the left camera intrinsic parameter matrix K l , the right camera intrinsic parameter matrix K r , the left camera distortion coefficient k li and the right camera distortion coefficient k ri (i=1,2), the external parameters include the rotation matrix R of the left camera relative to the world coordinate system wl With the translation vector t wl , the rotation matrix R of the right camera relative to the world coordinate system wr With the translation vector t wr The position relationship between the left camera and the right camera includes the rotation matrix R lr and the translation vector t lr , paste circular marking points on the surface of the cylindrical part, collect the side image of the cylindrical part through the binocular vision measurement device, and apply the left camera distortion coefficient k li and the right camera distortion coefficient k ri The collected side image of the cylinder is dedistorted.

3. The method for spatial posture tracking using a cylindrical marker eccentricity correction algorithm according to claim 2, characterized in that: The three-dimensional reconstruction and discrete sampling based on the collected image to generate the unfolded image after geometric distortion correction includes the following steps: Obtain the radius r of the cylindrical part, use the EDLines line detection algorithm to extract the cylindrical part contour lines l1 and l2 in the side image of the cylinder, and calculate the back-projection plane normal vector n corresponding to the contour line i : Solve for the bisecting plane normal vector n3: Among them, ||n i || represents the back-projected plane normal vector n i The modulus of the geometric parameters is then calculated, including the angle θ of the back-projection plane corresponding to the contour line, the axis direction vector s of the cylindrical part L , and the optical center O of the binocular vision measurement device C To the center P C The direction vector s OP : Determine the optical center O of the binocular vision measurement device C To the center P C The distance is: Calculate the center P of the circle C Coordinates: The three-dimensional model of the cylindrical part is completed.

4. The method for spatial posture tracking using a cylindrical marker eccentricity correction algorithm according to claim 3, characterized in that: The three-dimensional reconstruction and discrete sampling based on the acquired image to generate the unfolded image after geometric distortion correction also includes the following steps: In the camera coordinate system X C Y C Establish a standard circle C on the plane and discretize the central angle δ to obtain a point set Arc length The discretization process is performed along the axis of the cylindrical member to obtain the cross-sectional circle sequence {C j |j=1,2,…,J}, with the center of the cross-section circle P cj As the origin, the axis direction vector s of the cylindrical part L is the X axis, the bisecting plane normal vector n3 is the Y axis, and the direction vector s OP For the Z axis, establish a cross-section circle coordinate system, and use the transformation matrix between the binocular vision measurement device coordinate system and the cross-section circle coordinate system: Calculation of cross-section circle C j Discrete points on Coordinates on the standard circle C in Respectively and The homogeneous coordinates of .

5. The method for spatial posture tracking using a cylindrical marker eccentricity correction algorithm according to claim 4, characterized in that: The three-dimensional reconstruction and discrete sampling based on the acquired image to generate the unfolded image after geometric distortion correction also include the following steps: Calculate the discrete points described in claim 4 The optical center O of the binocular vision measurement device C The vector angle of : Where T1 is the contour tangent point. When β < 90°, it is determined to be a visible discrete point. The visible discrete point passes through: Get the coordinates of its projection point on the image in for The homogeneous coordinates, I is the identity matrix, and the cross-section circle C j The discrete point index i on the image is the row coordinate of the expanded image, and the index j is the column coordinate of the expanded image. Corresponding grayscale information Mapping to the corresponding position in the unfolded view matrix to generate an unfolded view of the outer surface of the cylindrical part.

6. The method for spatial posture tracking using a cylindrical marker eccentricity correction algorithm according to claim 5, characterized in that: The method of locating the center coordinates of the marker point by using an interpolation algorithm comprises the following steps: The EDCircles algorithm is used to detect the circular marker points in the flattened image and extract the initial circle center coordinates p fc =[u fc v fc ] T , locate the initial circle center coordinate p fc The four vertices of the pixel grid: Extract the decimal part of the coordinates: Compute exact circular coordinates via bilinear interpolation: What you want (u c ,v c ) are the center coordinates of the circular marking point on the outer surface of the cylindrical member.

7. The method for spatial posture tracking using a cylindrical marker eccentricity correction algorithm according to any one of claims 2 to 6, characterized in that: The method of solving the three-dimensional coordinates of the marker point by binocular triangulation and straight line linear transformation includes the following steps: By calculating and tracing the optical ray O emitted by the left camera l P l and the optical ray O emitted by the right camera r P r Intersection point P w , based on the linear transformation algorithm, the least squares estimation of the three-dimensional coordinates of the spatial point is completed: Among them, B is the coefficient matrix of the equation system, (u l ,v l ) and (u r ,v r ) are the intersection points P w Project point p on the left and right images l and p r The image coordinates of and is the projection matrix M l =K l [R wl t wl ] and M r =K r [R wr t wr ], and for and The transposed matrix of , solve the system of equations by linear least squares method Obtain P w An approximate solution to obtain the intersection point P w The three-dimensional coordinates in the camera coordinate system of the binocular vision measuring device are the position and posture of the cylindrical part in the coordinate system of the binocular vision measuring device.

8. A spatial posture tracking device using a cylindrical marker eccentricity correction algorithm, characterized in that: include: An image processing module includes a binocular vision measuring device, a tripod, and an arithmetic processor, wherein the binocular vision measuring device includes a left camera and a right camera, the tripod is connected to the bottom of the binocular vision measuring device, the binocular vision measuring device is used to capture an image of the cylindrical part to be measured, and the arithmetic processor is provided in the binocular vision measuring device to process and output required parameters; The marking point placement module includes a sliding platform, a cylindrical part, and circular marking points. Multiple circular marking points are distributed on the surface of the cylindrical part. The cylindrical part lies on its side above the sliding platform. The relative position of the cylindrical part and the binocular vision measurement device is adjusted by the sliding platform.

Images