Collaborative robot line structured light 3D vision hand-eye calibration method

Abstract

This invention relates to a collaborative robot line structured light 3D vision hand-eye calibration method, comprising: acquiring the robot's motion space; selecting fixed weld feature points in the robot's motion space; determining the coordinate values ​​of the fixed weld feature points in the robot's base coordinate system and the sensor coordinate system; acquiring the transformation relationship from the camera coordinate system to the robot's base coordinate system based on the coordinate values ​​of the fixed weld feature points in the robot's base coordinate system and the sensor coordinate system; acquiring the robot's posture expression mode; determining the transformation matrix between the robot's end effector coordinate system and the robot's base coordinate system, and the transformation matrix between the sensor coordinate system and the robot's end effector coordinate system, i.e., the hand-eye matrix, based on the robot's posture expression mode; and performing hand-eye calibration using the transformation relationship between the camera coordinate system and the robot's base coordinate system, the transformation matrix between the robot's end effector coordinate system and the robot's base coordinate system, and the hand-eye matrix.

Description

Technical Field

[0001] This invention relates to the field of robot 3D vision calibration technology, and in particular to a method for hand-eye calibration of collaborative robots using line structured light 3D vision. Background Technology

[0002] Currently, common hand-eye calibration methods include 3D target methods and 2D target methods. 2D targets mainly include checkerboard calibration boards and circular center calibration boards. During calibration, different poses of the target are acquired, and the hand-eye transformation matrix is ​​estimated by identifying corner points. There are many types of 3D targets. To accurately obtain the 3D coordinates of feature points, the 3D calibration sphere method is frequently used. The standard sphere method is based on a local arc formed by a laser line on a sphere, and calculates the sphere's center coordinates through a series of point cloud processing steps. Two existing, relatively mature calibration algorithms are introduced below:

[0003] One, Two-Step Method:

[0004] Tsai's two-step method is a classic hand-eye calibration method used to determine the conversion relationship between the robot's end effector and the camera. The main implementation of this method is as follows:

[0005] Step 1: Solve for the rotation relations:

[0006] The calibration board is placed still in the detection plane, the robot's pose is changed to perform a series of movements, and image data of the calibration board is collected at each pose.

[0007] The calibration board is used to obtain positioning information, and the rotation matrix and translation vector from the calibration board to the camera coordinate system are calculated from this information.

[0008] The camera's intrinsic and extrinsic parameters are obtained through camera calibration, and the relative rotation of the calibration plate between different robot poses is calculated from this data.

[0009] The rotation matrix in the hand-eye matrix is ​​calculated from the rotational differences between multiple sets of different poses using the Rodrigues formula or quaternion method.

[0010] Step 2: Solve for the translation relationship:

[0011] The influence of rotation components is removed using the rotation matrix obtained in the first step.

[0012] Using the rotation relation combined with the previously obtained relative translation, a system of linear equations is established.

[0013] The translation part of the hand-eye matrix is ​​solved by using the least squares method to solve the system of linear equations.

[0014] II. Auxiliary calibration ball method:

[0015] The calibration sphere method for hand-eye calibration primarily uses a calibration sphere with known geometric features to assist in positioning and determining the spatial transformation relationship between the robot's end effector and sensors, including cameras and line lasers. The calibration sphere method for hand-eye calibration involves the following steps:

[0016] Prepare a calibration ball and place it in the robot's workspace, so that it is within the camera's field of view. The laser emits a line laser that hits the surface of the calibration ball.

[0017] The robot's end effector moves the sensor to different positions and orientations of the calibration ball, recording the end effector pose matrix at each position; and a camera captures images of the calibration ball with a line laser on its surface.

[0018] The coordinates of the center of the tangent circle are obtained by analyzing the image of the calibration sphere acquired by the camera and by fitting the point cloud. Then, the coordinates of the center of the calibration sphere are solved.

[0019] By using multiple different robot end-effector pose matrices and their corresponding sphere center coordinates, a system of linear or nonlinear equations is constructed to solve for the transformation matrix of the robot end-effector relative to the camera coordinate system.

[0020] Based on the spatial relationship between the robot's end effector sensors and the relationship between the center of the sphere in the camera's field of view, the spatial transformation relationship between the robot's tool coordinate system and the camera coordinate system is solved, which is the result of hand-eye calibration.

[0021] Problems with the above-mentioned existing technical solutions:

[0022] Tsai's two-step hand-eye calibration requires the assistance of a checkerboard calibration board, and the calibration process is complicated, requiring multiple sets of calibration boards and corresponding robot poses. It also has high requirements for the accuracy of robot pose and calibration board pose data. Furthermore, the error of the rotation matrix in Tsai's two-step calibration process will accumulate in the solution of the translation vector.

[0023] The auxiliary calibration sphere method requires the use of an external calibration sphere, which is costly. The calculation of the sphere center by segmenting and fitting the point cloud of the tangential arc is easily affected by noise, and has high requirements for the imaging quality of the camera and the image processing algorithm. Due to the roughness of the standard sphere surface, not all points on the arc are on the same plane, which will cause the normal vector of the plane to shift, affecting the accuracy of the sphere center coordinates. Summary of the Invention

[0024] To address the problems existing in the prior art, the purpose of this invention is to propose a collaborative robot line structured light 3D vision hand-eye calibration method. From an application perspective, this method finds the optimal application form of the algorithm in practical applications, simplifies the hand-eye calibration steps, reduces the operational difficulty, and uses line structured light to assist in obtaining three-dimensional coordinates, thereby improving the accuracy of feature point recognition.

[0025] To achieve the above objectives, the present invention provides the following solution:

[0026] A collaborative robot line structured light 3D vision hand-eye calibration method includes:

[0027] Obtain the robot motion space, select a fixed weld feature point in the robot motion space, and determine the coordinate values ​​of the fixed weld feature point in the robot base coordinate system and the sensor coordinate system;

[0028] Based on the coordinate values ​​of the fixed weld feature points in the robot base coordinate system and the sensor coordinate system, the transformation relationship from the camera coordinate system to the robot base coordinate system is obtained.

[0029] Obtain the robot's posture expression mode, and based on the robot's posture expression mode, determine the transformation matrix between the robot's end-effector coordinate system and the robot's base coordinate system, and the transformation matrix between the sensor coordinate system and the robot's end-effector coordinate system, i.e., the hand-eye matrix;

[0030] Hand-eye calibration is performed using the transformation relationship between the camera coordinate system and the robot base coordinate system, the transformation matrix between the robot end effector coordinate system and the robot base coordinate system, and the hand-eye matrix.

[0031] Optionally, the method for obtaining the transformation relationship from the camera coordinate system to the robot base coordinate system is as follows:

[0032] P B =H TB H LT P L

[0033] Among them, P B Let P be the coordinates of point P in the robot's base coordinate system. L Let H be the coordinates of point P in space within the sensor coordinate system. TB H is the transformation matrix between the robot's end-effector coordinate system and the robot's base coordinate system. LT This is the transformation matrix between the sensor coordinate system and the robot end effector coordinate system, i.e., the hand-eye matrix.

[0034] Optionally, obtaining the robot's gesture expression method includes:

[0035] Based on Euler angles and quaternions, different pose representations are converted into pose matrices to obtain the robot's pose expression methods.

[0036] Optionally, determining the hand-eye matrix includes:

[0037]

[0038] Among them, H LTThe transformation matrix between the sensor coordinate system and the robot's end effector coordinate system, i.e., the hand-eye matrix, is r. 11 r 12 r 13 r 21 r 22 r 23 r 31 r 32 r 33 All of these are parameters of the rotation matrix in the hand-eye matrix, and t1, t2, and t3 are parameters of the translation vector in the hand-eye matrix.

[0039] Optionally, hand-eye calibration is performed using the transformation relationship from the camera coordinate system to the robot base coordinate system, the transformation matrix between the robot end effector coordinate system and the robot base coordinate system, and the hand-eye matrix, including:

[0040] Based on the hand-eye matrix and the transformation relationship from the camera coordinate system to the robot base coordinate system, the first transformation result is obtained:

[0041]

[0042] Based on the full rank of the transformation matrix between the robot end-effector coordinate system and the robot base coordinate system, the first transformation result is multiplied by H. TB -1 Obtain the second conversion result:

[0043]

[0044] The second conversion result is processed to obtain the third conversion result:

[0045]

[0046] Combine the second transformation result and the third transformation result to obtain the combined result:

[0047]

[0048] Where, x B y B z B All are the coordinates of the feature points in the robot's base coordinate system, H TB H is the transformation matrix from the robot's end-effector coordinate system to the robot's base coordinate system. TB -1 r is the inverse of the transformation matrix from the robot's end-effector coordinate system to the robot's base coordinate system. 11 r 12 r 13 r 21 r 22 r 23 r31 r 32 r 33 All are parameters of the rotation matrix in the hand-eye matrix, t1, t2, and t3 are parameters of the translation vector in the hand-eye matrix, and y L ,z L Here are the coordinates of the feature point in the sensor coordinate system, where X, Y, and Z are substitutes for H. TB -1 Left multiplication (x) B ,y B ,z B A matrix of ,1).

[0049] Optionally, hand-eye calibration using the transformation relationship from the camera coordinate system to the robot base coordinate system, the transformation matrix between the robot end effector coordinate system and the robot base coordinate system, and the hand-eye matrix further includes:

[0050] Establish a solution model using the combined results:

[0051]

[0052] Among them, X1, X2, X3, X4, Y1, Y2, Y3, Y4, Z1, Z2, Z3, and Z4 are all substitutes for H. TB -1 Left multiplication (x) B ,y B ,z B The numerical values ​​of the matrix r, 1) 11 r 12 r 13 r 21 r 22 r 23 r 31 r 32 r 33 All are parameters of the rotation matrix in the hand-eye matrix, t1, t2, and t3 are parameters of the translation vector in the hand-eye matrix, and y L1 y L2 y L3 y L4 z L1 z L2 z L3 z L4 These are all coordinate values ​​of a set of feature points in the sensor coordinate system;

[0053] The solution model is simplified, and the transpose is taken to obtain the transpose results:

[0054] D T X T =C T

[0055] Where D is four groups (X) i ,Y i Z i A matrix consisting of (1), where X is the hand-eye matrix and C is a matrix composed of four groups (0, Y). Li Z Li The matrix composed of ,1)

[0056] Set the solution conditions: A = D T B = C T X = X T ;

[0057] Based on the stated solution conditions, singular value decomposition is used to solve the transpose result to obtain the hand-eye matrix.

[0058] Optionally, hand-eye calibration using the transformation relationship from the camera coordinate system to the robot base coordinate system, the transformation matrix between the robot end effector coordinate system and the robot base coordinate system, and the hand-eye matrix further includes:

[0059] Define decomposition conditions, and use these conditions to perform singular value decomposition on the rotation matrix to be orthogonalized in the hand-eye matrix to obtain the rotation matrix:

[0060] R = UΛV T

[0061] Where R is the orthogonalized rotation matrix, U is the square matrix composed of the eigenvectors of P, Λ is the eigenma matrix, V is the square matrix composed of the eigenvectors of W, and T is the transpose sign.

[0062] The feature matrix is ​​replaced with the identity matrix E, and the standard orthogonal hand-eye matrix is ​​further obtained.

[0063] Optionally, the decomposition condition can be defined as:

[0064]

[0065] in, R is the rotation matrix in the hand-eye matrix. LT This is the transpose of the rotation matrix in the hand-eye matrix.

[0066] The beneficial effects of this invention are as follows:

[0067] This invention requires no additional calibration equipment, and the calibration process is simple. Simply select any weld feature point, move the robot to direct the laser line onto the feature point, and the sensor will recognize the feature point.

[0068] This invention uses line structured light to assist in the identification of three-dimensional feature points, resulting in high data acquisition accuracy.

[0069] This invention has low requirements for robot pose, is easy to acquire data, and requires little computation. By changing the robot pose, the hand-eye matrix can be calculated using only four sets of sensor data and the robot's pose itself.

[0070] The present invention has fewer calibration steps, is easy to operate, and introduces as few process errors as possible. Attached Figure Description

[0071] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0072] Figure 1 This is a flowchart of a collaborative robot line structured light 3D vision hand-eye calibration method according to an embodiment of the present invention;

[0073] Figure 2 This is a schematic diagram of a collaborative robot device according to an embodiment of the present invention. Detailed Implementation

[0074] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0075] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0076] like Figure 1 As shown in the figure, this embodiment discloses a collaborative robot line structured light 3D vision hand-eye calibration method, including: acquiring the robot motion space, selecting fixed weld feature points in the robot motion space, and determining the coordinate values ​​of the fixed weld feature points in the robot base coordinate system and the sensor coordinate system; based on the coordinate values ​​of the fixed weld feature points in the robot base coordinate system and the sensor coordinate system, acquiring the transformation relationship from the camera coordinate system to the robot base coordinate system; acquiring the robot's posture expression mode, and based on the robot's posture expression mode, determining the transformation matrix between the robot end effector coordinate system and the robot base coordinate system, and the transformation matrix between the sensor coordinate system and the robot end effector coordinate system (hand-eye matrix); and performing hand-eye calibration through the transformation relationship between the camera coordinate system and the robot base coordinate system, the transformation matrix between the robot end effector coordinate system and the robot base coordinate system, and the hand-eye matrix.

[0077] Specifically:

[0078] S1, Select any weld type, take a fixed weld point as the feature point, adjust the robot posture so that the robot welding torch is aligned with the feature point, and obtain the position of the feature point in the robot coordinate system;

[0079] S2, control the robot to move the line structured light sensor to the laser line to identify the feature point, and obtain the coordinates of the feature point in the sensor coordinate system; at the same time, record the robot end pose in the current posture;

[0080] S3, change the robot pose to ensure that the laser line recognizes the feature points, repeat step S2 four times, and obtain the feature point coordinates and their corresponding robot end poses under four different poses.

[0081] S4. Based on steps S1-S3, multiple sets of parameters are obtained, and the hand-eye calibration matrix is ​​calculated using calibration equation (1).

[0082] S5. Orthogonalize the hand-eye matrix from step S4 to obtain the final result.

[0083] like Figure 2 As shown, the collaborative robot device mainly includes a collaborative robot, a line structured light vision sensor, and a welding torch at the robot's end effector; the coordinate system is described in Table 1.

[0084] Table 1. Coordinate System Description

[0085]

[0086] Furthermore, the method for obtaining the transformation relationship from the camera coordinate system to the robot base coordinate system is as follows:

[0087] P B =H TB H LT P L

[0088] Among them, P B Let P be the coordinates of point P in the robot's base coordinate system. L Let H be the coordinates of point P in space within the sensor coordinate system. TB H is the transformation matrix between the robot's end-effector coordinate system and the robot's base coordinate system. LT This is the transformation matrix between the sensor coordinate system and the robot end effector coordinate system, i.e., the hand-eye matrix.

[0089] Furthermore, obtaining the robot's posture expression method includes: based on Euler angles and quaternions, determining different posture representations and converting them into posture matrices to obtain the robot's posture expression method.

[0090] Specifically, a fixed weld feature point P is selected in the robot's motion space. This point is in the robot's base coordinate system O. B The coordinates below are P B =(X B ,Y B Z B The coordinates of P in the sensor coordinate system are... L =(X L ,Y L Z L If point P is in the camera coordinate system, then the transformation relationship from the camera coordinate system to the robot base coordinate system is:

[0091] P B =H TB H LT P L (1)

[0092] In relation (1), P B P L This can be obtained through measurement. For the robot's end-effector pose matrix H... TB The matrix parameters can be obtained directly through the robot and can be represented in the following form:

[0093]

[0094] Where R and T represent the rotation matrix and translation vector, respectively:

[0095]

[0096] Robots typically represent position in Cartesian coordinates, but different robots express pose in different ways, using either Euler angles or quaternions. Both methods require conversion into a pose matrix for application. Below is an explanation of how different pose representations are converted into pose matrices:

[0097] Quaternion form:

[0098] Q=(wxyz) (4)

[0099] Among them, w, x, y, and z are the general expressions for quaternions.

[0100] According to equation (4), the transformation expression between rotation matrices and quaternions can be obtained:

[0101]

[0102] Euler angle form:

[0103] Because Euler angles can be combined in different ways, the way Euler angles represent orientation is not unique, and the final expression for the rotation matrix also varies. First, let's explain the different combinations and rotation methods of Euler angles:

[0104] There are six possible combinations of rotation sequences along the x, y, and z axes: (xyz, yzx, zxy, xzy, zyx, yxz). When rotating by the same angle, different rotation sequences result in different postures.

[0105] Assuming the rotation angles around the x, y, and z axes are α, β, and γ respectively, the rotation matrix for the three rotations is calculated as follows:

[0106]

[0107] Different manufacturers use different methods to express the rotation of robot postures, so the appropriate calculation method needs to be selected based on the posture rotation method of the robot being used.

[0108] Furthermore, determining the hand-eye matrix includes:

[0109]

[0110] Among them, H LT The transformation matrix between the sensor coordinate system and the robot's end effector coordinate system, i.e., the hand-eye matrix, is r. 11 r 12 r 13 r 21 r 22 r 23 r 31 r 32 r 33 All of these are parameters of the rotation matrix in the hand-eye matrix, and t1, t2, and t3 are parameters of the translation vector in the hand-eye matrix.

[0111] Furthermore, hand-eye calibration is performed using the transformation relationship from the camera coordinate system to the robot base coordinate system, the transformation matrix between the robot end effector coordinate system and the robot base coordinate system, and the hand-eye matrix, including:

[0112] Based on the hand-eye matrix and the transformation relationship from the camera coordinate system to the robot base coordinate system, obtain the first transformation result:

[0113]

[0114] Based on the full rank of the transformation matrix between the robot's end-effector coordinate system and the robot's base coordinate system, the first transformation result is multiplied by H. TB -1 Obtain the second conversion result:

[0115]

[0116] The second transformation result is processed to obtain the third transformation result:

[0117]

[0118] Combine the second and third transformation results to obtain the combined result:

[0119]

[0120] Where, x B y B z B All are the coordinates of the feature points in the robot's base coordinate system, H TB H is the transformation matrix from the robot's end-effector coordinate system to the robot's base coordinate system. TB -1 r is the inverse of the transformation matrix from the robot's end-effector coordinate system to the robot's base coordinate system. 11 r 12 r 13 r 21 r 22 r 23 r 31 r 32 r 33 All are parameters of the rotation matrix in the hand-eye matrix, t1, t2, and t3 are parameters of the translation vector in the hand-eye matrix, and y L ,z L Here are the coordinates of the feature point in the sensor coordinate system, where X, Y, and Z are substitutes for H. TB -1 Left multiplication (x) B ,y B ,z B A matrix of ,1).

[0121] Furthermore, hand-eye calibration, using the transformation relationship from the camera coordinate system to the robot base coordinate system, the transformation matrix between the robot end effector coordinate system and the robot base coordinate system, and the hand-eye matrix, also includes:

[0122] Establish a solution model using the results of simultaneous equations:

[0123]

[0124] Among them, X1, X2, X3, X4, Y1, Y2, Y3, Y4, Z1, Z2, Z3, and Z4 are all substitutes for H. TB -1 Left multiplication (x) B ,y B,z B The numerical values ​​of the matrix r, 1) 11 r 12 r 13 r 21 r 22 r 23 r 31 r 32 r 33 All are parameters of the rotation matrix in the hand-eye matrix, t1, t2, and t3 are parameters of the translation vector in the hand-eye matrix, and y L1 y L2 y L3 y L4 z L1 z L2 z L3 z L4 These are all coordinate values ​​of a set of feature points in the sensor coordinate system;

[0125] The solution model is simplified, and the transpose is taken to obtain the transpose results:

[0126] D T X T =C T

[0127] Where D is four groups (X) i ,Y i Z i A matrix consisting of (1), where X is the hand-eye matrix and C is a matrix composed of four groups (0, Y). Li Z Li The matrix composed of ,1)

[0128] Set the solution conditions: A = D T B = C T X = X T ;

[0129] Based on the solution conditions, singular value decomposition is used to solve the transpose result to obtain the hand-eye matrix.

[0130] Furthermore, hand-eye calibration, using the transformation relationship from the camera coordinate system to the robot base coordinate system, the transformation matrix between the robot end effector coordinate system and the robot base coordinate system, and the hand-eye matrix, also includes:

[0131] Define the decomposition conditions, and use these conditions to perform singular value decomposition on the rotation matrix to be orthogonalized in the hand-eye matrix to obtain the rotation matrix:

[0132] R = UΛV T

[0133] Where R is the orthogonalized rotation matrix, U is the square matrix composed of the eigenvectors of P, Λ is the eigenma matrix, V is the square matrix composed of the eigenvectors of W, and T is the transpose sign.

[0134] The feature matrix is ​​replaced with the identity matrix E, and the standard orthogonal hand-eye matrix is ​​further obtained.

[0135] Furthermore, the decomposition condition is defined as follows:

[0136]

[0137] in, R is the rotation matrix in the hand-eye matrix. LT This is the transpose of the rotation matrix in the hand-eye matrix.

[0138] Specifically,

[0139] The method described above can be used to determine H. TB All parameters for the hand-eye matrix H. LT Similarly, it can be expressed in the following form:

[0140]

[0141] From equations (1) and (7), we can obtain:

[0142]

[0143] Matrix H TB Full rank, multiply both sides of equation (8) by H on the left. TB -1 ,get:

[0144]

[0145] To process the left side of equation (9), let:

[0146]

[0147] Combining equations (9) and (10), we get:

[0148]

[0149] A group of P B ,P L The corresponding values ​​can be used to solve three equations. Since X has 12 parameters, at least 4 sets of corresponding values ​​are needed to solve for all 12 parameters.

[0150]

[0151] make:

[0152]

[0153] Equation (12) can be simplified to:

[0154] C = XD (13)

[0155] Taking the transpose of both sides of equation (13) gives:

[0156] C T =(XD) T (14)

[0157] Simplifying, we get:

[0158] D T X T =C T

[0159] Let A = D T B = C T X = X T Then equation (14) is:

[0160] AX = B (15)

[0161] Equation (15) is solved using Singular Value Decomposition (SVD), X T That is, the obtained hand-eye matrix H LT .

[0162] Equation (15) yields the numerical solution of the rotation matrix. Since the sensor coordinates lack X-axis data, an approximate rotation matrix R with errors is obtained. Orthogonality constraints are not considered, meaning the obtained rotation matrix R is not a standard orthogonal matrix and needs to be orthogonalized. Hand-eye matrix H LT Rotation matrix R LT for:

[0163]

[0164] definition For matrix R LT Perform singular value decomposition to obtain the rotation matrix:

[0165] R = UΛV T

[0166] The square matrix U is formed by the eigenvectors of P, the square matrix V is formed by the eigenvectors of W, and R is the orthogonalized rotation matrix. Replacing the eigenma matrix Λ with the identity matrix E yields the orthogonal matrix:

[0167] R = UEV T =UV T

[0168] At this time, the hand-eye matrix H LTThe standard orthogonal matrix is ​​used to complete the hand-eye calibration process.

[0169] The embodiments described above are merely preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Various modifications and improvements made to the technical solutions of the present invention by those skilled in the art without departing from the spirit of the present invention should fall within the protection scope defined by the claims of the present invention.

Claims

1. A method for hand-eye calibration of collaborative robots using line structured light 3D vision, characterized in that... include: Obtain the robot motion space, select a fixed weld feature point in the robot motion space, and determine the coordinate values ​​of the fixed weld feature point in the robot base coordinate system and the sensor coordinate system; Based on the coordinate values ​​of the fixed weld feature points in the robot base coordinate system and the sensor coordinate system, the transformation relationship from the camera coordinate system to the robot base coordinate system is obtained. Obtain the robot's posture expression mode, and based on the robot's posture expression mode, determine the transformation matrix between the robot's end-effector coordinate system and the robot's base coordinate system, and the transformation matrix between the sensor coordinate system and the robot's end-effector coordinate system, i.e., the hand-eye matrix; Determining the hand-eye matrix includes: Among them, H LT The transformation matrix between the sensor coordinate system and the robot's end effector coordinate system, i.e., the hand-eye matrix, is r. 11 r 12 r 13 r 21 r 22 r 23 r 31 r 32 r 33 All of these are parameters of the rotation matrix in the hand-eye matrix, and t1, t2, and t3 are parameters of the translation vector in the hand-eye matrix; Hand-eye calibration is performed using the transformation relationship from the camera coordinate system to the robot base coordinate system, the transformation matrix between the robot end effector coordinate system and the robot base coordinate system, and the hand-eye matrix. Based on the hand-eye matrix and the transformation relationship from the camera coordinate system to the robot base coordinate system, the first transformation result is obtained: Based on the full rank of the transformation matrix between the robot end-effector coordinate system and the robot base coordinate system, the first transformation result is multiplied by H. TB -1 Obtain the second conversion result: The second conversion result is processed to obtain the third conversion result: Combine the second transformation result and the third transformation result to obtain the combined result: in, , , All are the coordinates of the feature points in the robot's base coordinate system, H TB H is the transformation matrix from the robot's end-effector coordinate system to the robot's base coordinate system. TB -1 r is the inverse of the transformation matrix from the robot's end-effector coordinate system to the robot's base coordinate system. 11 r 12 r 13 r 21 r 22 r 23 r 31 r 32 r 33 All of these are parameters of the rotation matrix in the hand-eye matrix, and t1, t2, and t3 are parameters of the translation vector in the hand-eye matrix. , These are the coordinates of the feature point in the sensor coordinate system. , , All are substitutes for H TB -1 Left multiplication (x) B ,y B ,z B A matrix of ,1).

2. The collaborative robot line structured light 3D vision hand-eye calibration method according to claim 1, characterized in that, The method for obtaining the transformation relationship from the camera coordinate system to the robot base coordinate system is as follows: Among them, P B Let P be the coordinates of point P in the robot's base coordinate system. L Let H be the coordinates of point P in space within the sensor coordinate system. TB H is the transformation matrix between the robot's end-effector coordinate system and the robot's base coordinate system. LT This is the transformation matrix between the sensor coordinate system and the robot end effector coordinate system, i.e., the hand-eye matrix.

3. The collaborative robot line structured light 3D vision hand-eye calibration method according to claim 1, characterized in that, The methods for obtaining the robot's pose expression include: Based on Euler angles and quaternions, different pose representations are converted into pose matrices to obtain the robot's pose expression methods.

4. The collaborative robot line structured light 3D vision hand-eye calibration method according to claim 1, characterized in that, Hand-eye calibration, which utilizes the transformation relationship from the camera coordinate system to the robot base coordinate system, the transformation matrix between the robot end effector coordinate system and the robot base coordinate system, and the hand-eye matrix, also includes: A solution model is established using the combined results: in, , , , , , , , , , , , All are substitutes for H TB -1 Left multiplication (x) B ,y B ,z B The numerical values ​​of the matrix r, 1) 11 r 12 r 13 r 21 r 22 r 23 r 31 r 32 r 33 All of these are parameters of the rotation matrix in the hand-eye matrix, and t1, t2, and t3 are parameters of the translation vector in the hand-eye matrix. , , , , , , , These are all coordinate values ​​of a set of feature points in the sensor coordinate system; The solution model is simplified, and the transpose is taken to obtain the transpose results: D T X T =C T Where D is four groups (X) i ,Y i Z i A matrix consisting of (1), where X is the hand-eye matrix and C is a matrix composed of four groups (0, Y). Li Z Li The matrix composed of ,1) Set the solution condition: A=D T B=C T X=X T ; Based on the stated solution conditions, singular value decomposition is used to solve the transpose result to obtain the hand-eye matrix.

5. The collaborative robot line structured light 3D vision hand-eye calibration method according to claim 4, characterized in that, Hand-eye calibration, which utilizes the transformation relationship from the camera coordinate system to the robot base coordinate system, the transformation matrix between the robot end effector coordinate system and the robot base coordinate system, and the hand-eye matrix, also includes: Define decomposition conditions, and use these conditions to perform singular value decomposition on the rotation matrix to be orthogonalized in the hand-eye matrix to obtain the rotation matrix: R=UΛV T Where R is the orthogonalized rotation matrix, U is the square matrix composed of the eigenvectors of P, Λ is the eigenma matrix, V is the square matrix composed of the eigenvectors of W, and T is the transpose sign. The feature matrix is ​​replaced with the identity matrix E, and the standard orthogonal hand-eye matrix is ​​further obtained.

6. The collaborative robot line structured light 3D vision hand-eye calibration method according to claim 5, characterized in that, The decomposition condition is defined as follows: , in, This is the rotation matrix in the hand-eye matrix. This is the transpose of the rotation matrix in the hand-eye matrix.

Images