CN117601126A

CN117601126A - Method, equipment and medium for detecting singular resolution of robot hand-eye calibration analysis

Info

Publication number: CN117601126A
Application number: CN202311686380.4A
Authority: CN
Inventors: 张翼翔; 曹安港; 王聚中; 薛清海; 吴强强; 甘华强; 宋文峰; 张艳超; 刘红松
Original assignee: Zhengzhou Dongxing Environmental Protection Energy Co ltd
Current assignee: Zhengzhou Dongxing Environmental Protection Energy Co ltd
Priority date: 2023-12-08
Filing date: 2023-12-08
Publication date: 2024-02-27

Abstract

The application relates to the technical field of robot vision calibration, in particular to a method, equipment and medium for detecting robot hand-eye calibration analysis solution singularity. The method comprises the following steps: acquiring visual image information and mechanical arm information, and setting true values of a hand-eye matrix ⁿ x, three groups of semantic measurement data are screened out according to visual image information and mechanical arm information, actual measurement data with noise conditions are determined, measurement data extraction is carried out on the actual measurement data, an actual measurement data rotation axis vector, a rotation angle value and a position column vector are obtained, solution analysis is carried out based on parameter analysis logic, a hand eye matrix estimated value is obtained, the hand eye matrix estimated value is analyzed, angle error values are obtained, and three groups of actual measurement data and true values are based on the three groups of actual measurement data ⁿ And x and the angle error value determine a singular data group, and carry out measurement data deviation correction on the singular data group according to the angle error value, thereby improving the safety of the hand-eye calibration algorithm.

Description

Method, equipment and medium for detecting singular resolution of robot hand-eye calibration analysis

Technical Field

The application relates to the technical field of robot vision calibration, in particular to a method, equipment and medium for detecting resolution singularity of robot hand-eye calibration.

Background

In recent years, with the rapid development of science and technology, robots have gradually entered our social life, such as being widely used in the fields of intelligent assembly, autonomous navigation, reverse engineering, welding engineering and the like, the vision-guided robot technology is increasingly applied to efficient high-quality processing, the system calibration precision is a primary factor for ensuring the processing quality, and the hand-eye calibration is a key step in the calibration process.

Currently, solving a transition matrix between a robot tool coordinate system (Hand) and a clamped camera coordinate system (Eye) is a key process for completing a visual operation task for Eye-in-Hand type robot Hand-Eye visual systems, and the process is a so-called Hand-Eye calibration process. At present, the numerical solution precision and the calculation complexity of the hand-eye calibration process are greatly dependent on given initial conditions, and according to the parameterized difference of the rotation matrix, the hand-eye calibration analysis solution mainly comprises a solution based on a directional cosine matrix (Direction Cosine Matrix), a closed solution based on an Euler Axis Angle (Euler Axis-Angle) parameter, a closed solution based on a modified Rodeux (Modified Rodrigues) parameter, a closed solution based on a Quaternion (Quaternion Algebra) parameter, a closed solution based on a Euclidean Group (Euclidean Group) parameter, a closed solution based on a Dual Quaternion/spiral theory (Dual Quaternion) parameter and a closed solution based on an orthogonal Dual tensor (Orthogonal Dual Tensor) parameter. In view of the difficulty in ensuring that the obtained unknown hand-eye matrix posture part is always an orthogonal matrix with determinant equal to +1 in the hand-eye calibration method based on the direction cosine matrix parameters, the hand-eye calibration process is analyzed by the hand-eye calibration analysis method based on the remaining six parameters.

The method is characterized in that the method can be obtained on the basis of multiple test results of one or more hand-eye calibration analysis solutions in the remaining six types, wherein the one or more hand-eye calibration analysis solutions have completely wrong unknown hand-eye matrix estimated values, and the corresponding hand-eye calibration method is invalid at the moment, namely, the hand-eye calibration singular phenomenon, which directly leads to the failure of the robot vision operation task. From the safety aspect, whether the hand-eye calibration analysis solution can avoid the singular phenomenon under some specific conditions is still a key for evaluating whether the robot hand-eye calibration is successful or not. Therefore, it is particularly important to analyze the singularities of the existing hand-eye calibration analysis solutions, so that not only can the trigger conditions of the singularities be given, but also the safety of the hand-eye calibration algorithm can be improved to a certain extent.

Disclosure of Invention

In order to solve at least one technical problem, the application provides a method, a device, electronic equipment and a medium for detecting the resolution singularity of the calibration and analysis of the hands and eyes of a robot.

In a first aspect, the present application provides a method for detecting singular resolution of calibration analysis of a robot hand and an eye, which adopts the following technical scheme:

a control method of an electronic device, comprising:

visual image information and mechanical arm information are acquired, wherein the visual image information is used for representing visual image information shot by a four-eye camera arranged on a mechanical arm of a robot in a history period, and the mechanical arm information is used for representing motion information of the mechanical arm of the robot in the history period;

Setting true value of hand-eye matrix ⁿ x, saidWherein, ⁿ t represents a true value ⁿ x mechanical arm position column vector, rot @ ⁿ k, ⁿ θ _x ) Representing true values ⁿ The robot arm of x rotates the data, ⁿ θ _x is Rot% ⁿ k, ⁿ θ _x ) The upper left hand sign n is the abbreviation of nominal for calibrating nominal data;

three groups of nominal measurement data { of non-parallel rotating shafts of the mechanical arm under the noiseless condition are screened out according to the visual image information and the mechanical arm information ⁿ A _i , ⁿ B _i -wherein i = 1,2,3 and the rotation angle value of the nominal measurement data when i = 3 is constrained to be equal to pi radians;

gauss noise is added to the three groups of nominal measurement data to obtain actual measurement data { corresponding to the three groups of nominal measurement data under the noise condition respectively ^a A _i , ^a B _i -wherein the upper left hand mark a is an abbreviation for actual for calibrating the actual measurement data;

respectively extracting measurement data of three groups of actual measurement data to obtain a rotation axis vector, a rotation angle value and a position column vector corresponding to each group of actual measurement data;

solving and analyzing the rotation axis vector, the rotation angle value and the position column vector corresponding to each group of actual measurement data based on different parameter analysis logics to obtain hand-eye matrix estimated values of different combinations of actual measurement data corresponding to different parameter analysis; performing precision evaluation analysis on the hand-eye matrix estimated values of different combinations of actual measurement data corresponding to different parameter analyses to obtain an angle error value;

Based on three sets of the actual measurement data { ^a A _i , ^a B _i Go and true value ⁿ x and the angle error value determine three sets of the actual measurement data { ^a A _i , ^a B _i And calibrating a singular data group of the singular phenomenon by the hand and the eye, and carrying out measurement data deviation correction on the singular data group according to the angle error value.

The present application may be further configured in a preferred example to: and respectively extracting the measurement data of the three groups of actual measurement data to obtain a rotation axis vector, a rotation angle value and a position column vector corresponding to each group of actual measurement data, and then further comprising:

judging each group of practical test data { according to the rotation angle value ^a A _i , ^a B _i In } ^a A _i Is a rotation matrix R of (2) _Ai The corresponding rotation angle value is { with the actual test data ^a A _i , ^a B _i In } ^a B _i Is a rotation matrix R of (2) _Bi Whether the rotation angle difference value of the corresponding rotation angle value accords with a preset difference value range or not;

if the rotation angle difference accords with the preset difference range, the rotation matrix R is obtained _Ai And the rotation angle value of the rotation matrix R _Bi The rotation angle value of (2) is limited to a range of (0, pi) or (pi, 2 pi).

The present application may be further configured in a preferred example to: the method for solving and analyzing the rotation axis vector, the rotation angle value and the position column vector corresponding to each group of actual measurement data based on different parameter analysis logic to obtain hand-eye matrix estimated values of different combinations of actual measurement data corresponding to different parameter analysis comprises the following steps:

Converting the rotation axis vector, the rotation angle value and the position column vector into Euler axis angle parameters, substituting the Euler axis angle parameters into a hand-eye calibration analysis solution of the Euler axis angle parameters to obtain the first two groups of actual measurement data { ^a A _i , ^a B _i Hand-eye matrix estimation value corresponding to (i=1, 2)Three sets of actual measurement data { ^a A _i , ^a B _i Hand-eye matrix estimation value corresponding to (i=1, 2, 3)>

Converting the rotation axis vector, the rotation angle value and the position column vector into modified Rodrigas parameters, and substituting the modified Rodrigas parameters into a hand-eye calibration analysis solution for modifying the Rodrigas parameters to obtain the first two groups of actual measurement data { ^a A _i , ^a B _i Hand-eye matrix estimation value corresponding to (i=1, 2)Three sets of actual measurement data { ^a A _i , ^a B _i Hand-eye matrix estimation value corresponding to (i=1, 2, 3)>

Converting the rotation axis vector, the rotation angle value and the position column vector into quaternion parameters, and substituting the quaternion parameters into a hand-eye calibration analysis solution of the quaternion parameters to obtain the first two groups of actual measurement data { ^a A _i , ^a B _i Hand-eye matrix estimation value corresponding to (i=1, 2)Three sets of actual measurement data { ^a A _i , ^a B _i Hand-eye matrix estimation value corresponding to (i=1, 2, 3)

Converting the rotation axis vector, the rotation angle value and the position column vector into Euclidean group parameters, substituting the Euclidean group parameters into a hand-eye calibration analysis solution of the Euclidean group parameters to obtain the first two groups of actual measurement data { ^a A _i , ^a B _i Hand-eye matrix estimation value corresponding to (i=1, 2)Three sets of actual measurement data { ^a A _i , ^a B _i Hand-eye matrix estimation value corresponding to (i=1, 2, 3)>

Converting the rotation axis vector, rotation angle value and position column vector into dual quaternion/spiral theoretical parameters, and substituting the dual quaternion/spiral theoretical parameters into dualIn the hand-eye calibration analysis solution of quaternion/spiral theoretical parameters, the front two groups of actual measurement data { are obtained ^a A _i , ^a B _i Hand-eye matrix estimation value corresponding to (i=1, 2)Three sets of actual measurement data { ^a A _i , ^a B _i Hand-eye matrix estimation value corresponding to (i=1, 2, 3)>

Converting the rotation axis vector, the rotation angle value and the position column vector into orthogonal dual tensor parameters, substituting the orthogonal dual tensor parameters into a hand-eye calibration analysis solution of the orthogonal dual tensor parameters to obtain the first two groups of actual measurement data { ^a A _i , ^a B _i Hand-eye matrix estimation value corresponding to (i=1, 2)Three sets of actual measurement data { ^a A _i , ^a B _i Hand-eye matrix estimation value corresponding to (i=1, 2, 3)>

The present application may be further configured in a preferred example to: the performing precision evaluation analysis on the hand-eye matrix estimation values of different combinations of actual measurement data corresponding to different parameter analyses to obtain an angle error value comprises the following steps:

obtaining an angular distance error criterion d _∠ The angular distance error criterion d _∠ The rotation matrix precision for measuring the hand-eye matrix estimated value corresponding to the first two groups of actual measurement data and the three groups of actual measurement data of the different parameter analysis is measured, the rotation matrix precision is that

Wherein the logm is a matrix logarithm operator, theFor the rotation data of the mechanical arm in the matrix estimation value of the hand and eye, the following is carried out ⁿ R _x Is true value ⁿ Manipulator rotation data in matrix estimation value corresponding to x, II _F Is the Frobenius norm operator, the theta _δ Reading a rotation angle;

substituting the hand-eye matrix estimated values of different combinations of actual measurement data corresponding to the different parameter analyses intoSaid true value ⁿ Substituting the rotation data of the mechanical arm in x into ⁿ R _x And calculating to obtain angle error values corresponding to different combinations of actual measurement data corresponding to different parameter analyses.

The present application may be further configured in a preferred example to: the three groups of nominal measurement data { of non-parallel mechanical arm rotation axes under the noiseless condition are screened out according to the visual image information and the mechanical arm information ⁿ A _i , ⁿ B _i -comprising:

determining actual measurement data { according to the visual image information and the mechanical arm information ^a A _i , ^a B _i Homogeneous transformation matrix when i=1 in }Pose matrix->Actual measurement data { ^a A _i , ^a B _i 1.ltoreq.i in }<Homogeneous transformation matrix at 3->Pose matrix->Actual measurement data { ^a A _i , ^a B _i Homogeneous transformation matrix +.f when i=3 in }>Pose matrix->Wherein the homogeneous transformation matrix ∈ ->In order to change the mechanical tool system from the initial state to the first configuration in the visual range of the four-eye camera, the mechanical tool system is relative to the homogeneous transformation matrix of the basic system, and the homogeneous transformation matrix is>In order to change the mechanical tool system from the first configuration to the second configuration to the homogeneous transformation matrix of the basic system in the visual range of the four-eye camera, the homogeneous transformation matrix>In order to be in the visual range of the four-eye camera, the mechanical tool system after the mechanical arm is transformed from the second configuration to the third configuration is a homogeneous transformation matrix relative to the basic system;

wherein the pose matrixIn order to obtain a position and posture matrix of the calibration plate under a camera coordinate system in the calibration plate image shot by the four-eye camera after the mechanical arm is transformed to the first configuration, the posture matrix is>To the four-eye camera after the mechanical arm is converted to the second configuration Position and posture matrix of the calibration plate under the camera coordinate system in the photographed calibration plate image, wherein the posture matrix is +.>The position and posture matrix of the calibration plate under the camera coordinate system in the calibration plate image shot by the four-eye camera after the mechanical arm is transformed to the third configuration;

calculating homogeneous transformation matrix in two adjacent movements to obtain transition matrix

Calculating the pose matrix between two adjacent motions to obtain a relative pose matrix

Judging whether j in the transition matrix and the relative pose matrix is equal to a preset value, if not, increasing the value of j (j=j+1), and circularly executing the steps until j is equal to the preset value;

acquiring the acquisition times of data corresponding to the transition matrix and the relative pose matrix, judging whether the acquisition times exceed preset acquisition times, if not, repeating the steps until the data acquisition times exceed the preset acquisition times, if so, determining the transition matrix data and the relative pose matrix data corresponding to the acquisition times each time, and screening the transition matrix data and the relative pose matrix data according to preset screening conditions to obtain data with non-parallel rotating shafts of three groups of mechanical arms.

In a second aspect, the present application provides a device for detecting singular resolution of calibration and analysis of a robot hand and eye, which adopts the following technical scheme:

a device for detecting the resolution singularity of a robot hand-eye calibration analysis, which comprises,

the information acquisition module is used for acquiring visual image information and mechanical arm information, wherein the visual image information is used for representing visual image information shot by a four-eye camera arranged on the mechanical arm of the robot in a history period, and the mechanical arm information is used for representing movement information of the mechanical arm of the robot in the history period;

the truth value setting module is used for setting the truth value of the hand-eye matrix ⁿ x, saidWherein, ⁿ t represents a true value ⁿ x mechanical arm position column vector, rot @ ⁿ k, ⁿ θ _x ) Representing true values ⁿ The robot arm of x rotates the data, ⁿ θ _x is Rot% ⁿ k, ⁿ θ _x ) The upper left hand sign n is the abbreviation of nominal for calibrating nominal data;

the data screening module is used for screening three groups of nominal measurement data { of non-parallel mechanical arm rotation axes under the noiseless condition according to the visual image information and the mechanical arm information ⁿ A _i , ⁿ B _i -wherein i = 1,2,3 and the rotation angle value of the nominal measurement data when i = 3 is constrained to be equal to pi radians;

The noise mixing module is used for adding Gauss noise to the three groups of nominal measurement data to obtain actual measurement data { corresponding to the three groups of nominal measurement data under the noise condition respectively ^a A _i , ^a B _i -wherein the upper left hand mark a is an abbreviation for actual for calibrating the actual measurement data;

the data extraction module is used for extracting measurement data of three groups of actual measurement data respectively to obtain a rotation axis vector, a rotation angle value and a position column vector corresponding to each group of actual measurement data;

the data analysis module is used for carrying out solution analysis on the rotation axis vector, the rotation angle value and the position column vector corresponding to each group of actual measurement data based on different parameter analysis logics to obtain hand-eye matrix estimated values of different combinations of actual measurement data corresponding to different parameter analysis;

the error analysis module is used for carrying out precision evaluation analysis on the hand-eye matrix estimated values of different combinations of actual measurement data corresponding to different parameter analyses to obtain an angle error value;

a singular determination module for { based on three sets of said actual measurement data ^a A _i , ^a B _i Go and true value ⁿ x and the angle error value determine three sets of the actual measurement data { ^a A _i , ^a B _i And calibrating a singular data group of the singular phenomenon by the hand and the eye, and carrying out measurement data deviation correction on the singular data group according to the angle error value.

In one possible implementation, the apparatus further includes: the difference judging module and the angle value limiting module, wherein,

the difference judging module is used for judging each group of actual test data { according to the rotation angle value ^a A _i , ^a B _i In } ^a A _i Is a rotation matrix R of (2) _Ai The corresponding rotation angle value is { with the actual test data ^a A _i , ^a B _i In } ^a B _i Is a rotation matrix R of (2) _Bi Whether the rotation angle difference value of the corresponding rotation angle value accords with a preset difference value range or not;

the angle value limiting module is used for limiting the rotation matrix R when the rotation angle difference value accords with a preset difference value range _Ai And the rotation angle value of the rotation matrix R _Bi The rotation angle value of (2) is limited to a range of (0, pi) or (pi, 2 pi).

In another possible implementation manner, when the data analysis module performs solution analysis on the rotation axis vector, the rotation angle value and the position column vector corresponding to each set of actual measurement data based on different parameter analysis logic, the data analysis module is specifically configured to:

converting the rotation axis vector, the rotation angle value and the position column vector into Euler axis angle parameters, substituting the Euler axis angle parameters into a hand-eye calibration analysis solution of the Euler axis angle parameters to obtain the first two groups of actual measurement data { ^a A _i , ^a B _i Is } (i=1, 2)Corresponding hand-eye matrix estimation valueThree sets of actual measurement data { ^a A _i , ^a B _i Hand-eye matrix estimation value corresponding to (i=1, 2, 3)>

Converting the rotation axis vector, the rotation angle value and the position column vector into Euclidean group parameters, substituting the Euclidean group parameters into hand-eye calibration analysis solutions of the Euclidean group parameters,obtain the two previous sets of actual measurement data { ^a A _i , ^a B _i Hand-eye matrix estimation value corresponding to (i=1, 2)Three sets of actual measurement data { ^a A _i , ^a B _i Hand-eye matrix estimation value corresponding to (i=1, 2, 3)>

Converting the rotation axis vector, the rotation angle value and the position column vector into dual quaternion/spiral theoretical parameters, substituting the dual quaternion/spiral theoretical parameters into a hand-eye calibration analysis solution of the dual quaternion/spiral theoretical parameters to obtain the first two groups of actual measurement data { ^a A _i , ^a B _i Hand-eye matrix estimation value corresponding to (i=1, 2)Three sets of actual measurement data { ^a A _i , ^a B _i Hand-eye matrix estimation value corresponding to (i=1, 2, 3)>

In another possible implementation manner, the error analysis module is specifically configured to, when performing accuracy evaluation analysis on the hand-eye matrix estimated values of different combinations of actual measurement data corresponding to different parameter analyses to obtain an angle error value:

In another possible implementation manner, the data filtering module filters three groups of nominal measurement data { of non-parallel rotation axes of the mechanical arm under the noiseless condition according to the visual image information and the mechanical arm information ⁿ A _i , ⁿ B _i When } is used, in particular:

according to the describedVisual image information and the mechanical arm information determine actual measurement data { ^a A _i , ^a B _i Homogeneous transformation matrix when i=1 in }Pose matrix->Actual measurement data { ^a A _i , ^a B _i 1.ltoreq.i in }<Homogeneous transformation matrix at 3->Pose matrix->Actual measurement data { ^a A _i , ^a B _i Homogeneous transformation matrix +.f when i=3 in } >Pose matrix->Wherein the homogeneous transformation matrix ∈ ->In order to change the mechanical tool system from the initial state to the first configuration in the visual range of the four-eye camera, the mechanical tool system is relative to the homogeneous transformation matrix of the basic system, and the homogeneous transformation matrix is>In order to change the mechanical tool system from the first configuration to the second configuration to the homogeneous transformation matrix of the basic system in the visual range of the four-eye camera, the homogeneous transformation matrix>To be at the fourIn the visual range of the eye camera, the mechanical arm is transformed from the second configuration to the homogeneous transformation matrix of the mechanical tool system relative to the basic system after the third configuration;

wherein the pose matrixIn order to obtain a position and posture matrix of the calibration plate under a camera coordinate system in the calibration plate image shot by the four-eye camera after the mechanical arm is transformed to the first configuration, the posture matrix is>In order to obtain a position and posture matrix of the calibration plate under a camera coordinate system in the calibration plate image shot by the four-eye camera after the mechanical arm is transformed to the second configuration, the posture matrix is>The position and posture matrix of the calibration plate under the camera coordinate system in the calibration plate image shot by the four-eye camera after the mechanical arm is transformed to the third configuration;

In a third aspect, the present application provides an electronic device, which adopts the following technical scheme:

at least one processor;

a memory;

at least one application program, wherein the at least one application program is stored in the memory and configured to be executed by the at least one processor, the at least one application program configured to: the method for detecting the singular resolution of the calibration and analysis of the robot hand and eye is executed.

In a fourth aspect, the present application provides a computer readable storage medium, which adopts the following technical scheme:

a computer-readable storage medium having stored thereon a computer program which, when executed in a computer, causes the computer to perform a method of detecting robotic hand-eye calibration resolution singularities as described above.

In summary, the present application includes at least one of the following beneficial technical effects:

by adopting the technical scheme, when the measured data deviation correction is carried out on the singular data generated by hand eye calibration, visual image information and mechanical arm information are obtained, wherein the visual image information is used for representing visual image information shot in a history period by a four-eye camera arranged on the mechanical arm of the robot, the mechanical arm information is used for representing movement information of the mechanical arm of the robot in the history period, the measured data of the mechanical arm is collected according to the visual image information and the mechanical arm information, and then a true value of a hand eye matrix is set ⁿ x is a nominal hand-eye calibration analysis value corresponding to the manipulator position column vector and the manipulator rotation data, and three groups of manipulator rotation under the noiseless condition are selected according to the visual image information and the manipulator information Nominal measurement data { with non-parallel axes ⁿ A _i , ⁿ B _i Then Gauss noise is added to the three groups of sense measurement data to obtain actual measurement data { corresponding to the three groups of sense measurement data under the noise condition respectively ^a A _i , ^a B _i Because of a certain noise interference in the normal hand-eye calibration analysis process, adding noise into nominal measurement data with non-parallel rotating shafts of a mechanical arm to obtain actual measurement data, improving the authenticity of the measurement data, respectively extracting measurement data from three groups of actual measurement data to obtain rotating shaft vectors, rotating angle values and position column vectors corresponding to each group of actual measurement data, solving and analyzing the rotating shaft vectors, the rotating angle values and the position column vectors corresponding to each group of actual measurement data according to different parameter analysis logics mentioned in the background art to obtain hand-eye matrix estimated values of different combinations of actual measurement data corresponding to different parameter analysis, performing precision evaluation analysis on hand-eye matrix estimated values of different combinations of actual measurement data corresponding to different parameter analysis to obtain angle error values, and then carrying out { based on the three groups of actual measurement data ^a A _i , ^a B _i Go and true value ⁿ x and angle error values determine three sets of actual measurement data { ^a A _i , ^a B _i And the singular data set of the hand-eye calibration singular phenomenon exists in the step, and the measured data deviation correction is carried out on the singular data set according to the angle error value, so that the measured data in the singular data set is corrected in a deviation mode while the hand-eye calibration singular phenomenon is verified, and the safety of a hand-eye calibration algorithm is improved.

Drawings

Fig. 1 is a schematic flow chart of a method for detecting resolution singularities of robot hand-eye calibration analysis provided in an embodiment of the present application;

fig. 2 is a schematic diagram of a manipulator arm and hand calibration device for detecting a checkerboard calibration plate of a method for resolving singularities by calibrating and resolving a manipulator and hand according to an embodiment of the present application;

fig. 3 is a pose recording logic block diagram of a method for detecting resolution singularities of robot hand-eye calibration analysis according to an embodiment of the present application;

fig. 4 is a logic block diagram of solving and screening data errors of a method for detecting singular solutions of calibration and analysis of hands and eyes of a robot according to an embodiment of the present application;

fig. 5 is a schematic diagram of an error-invariant numerical experiment given by a hand-eye calibration solution of the method for detecting singularity of a hand-eye calibration solution of a robot according to an embodiment of the present application;

fig. 6 is a schematic structural diagram of a device for detecting resolution singularities of calibration and analysis of hands and eyes of a robot according to an embodiment of the present application;

Fig. 7 is a schematic structural diagram of an electronic device according to an embodiment of the present application.

Detailed Description

The present application is described in further detail below with reference to fig. 1 to 7.

The present embodiment is merely illustrative of the present application and is not intended to be limiting, and those skilled in the art, after having read the present specification, may make modifications to the present embodiment without creative contribution as required, but is protected by patent laws within the scope of the claims of the present application.

For the purposes of making the objects, technical solutions and advantages of the embodiments of the present application more clear, the technical solutions of the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is apparent that the described embodiments are some embodiments of the present application, but not all embodiments. All other embodiments, which can be made by one of ordinary skill in the art based on the embodiments herein without making any inventive effort, are intended to be within the scope of the present application.

In addition, the term "and/or" herein is merely an association relationship describing an association object, and means that three relationships may exist, for example, a and/or B may mean: a exists alone, A and B exist together, and B exists alone. In this context, unless otherwise specified, the term "/" generally indicates that the associated object is an "or" relationship.

Embodiments of the present application are described in further detail below with reference to the drawings attached hereto.

The embodiment of the application provides a method for detecting the resolution singularity of robot hand-eye calibration, which is executed by electronic equipment, wherein the electronic equipment can be a server or terminal equipment, and the server can be an independent physical server, a server cluster or a distributed system formed by a plurality of physical servers, or a cloud server for providing cloud computing service. The terminal device may be, but is not limited to, a smart phone, a tablet computer, a notebook computer, a desktop computer, etc., and the terminal device and the server may be directly or indirectly connected through wired or wireless communication, which is not limited herein, and as shown in fig. 1, the method includes:

step S10: visual image information and mechanical arm information are acquired.

The visual image information is used for representing visual image information shot by a four-eye camera arranged on a robot arm in a history period, and the robot arm information is used for representing movement information of the robot arm in the history period.

As shown in fig. 2, a checkerboard calibration plate is placed at the horizontal surface, a robot manipulator is arranged on one side of the checkerboard calibration plate, a four-eye camera is arranged right above the robot manipulator, and the robot manipulator performs torsion movement according to a corresponding shooting rule, so that the four-eye camera can shoot corresponding checkerboard calibration plate image information, namely visual image information, and torsion movement information of the robot manipulator is the manipulator information.

Step S11: setting true value of hand-eye matrix ⁿ x，

Wherein, ⁿ t represents a true value ⁿ x mechanical arm position column vector, rot @ ⁿ k, ⁿ θ _x ) Representing true values ⁿ The robot arm of x rotates the data, ⁿ θ _x is Rot% ⁿ k, ⁿ θ _x ) The upper left hand sign n is an abbreviation for nominal, used to calibrate the nominal data.

Specifically, truth values of hand-eye matrix ⁿ x represents the data obtained under noise-free conditions, such data being defined as nominal data in the examples of the present application. True value ⁿ The data parameters contained in x are real movement data parameters of the robot manipulator, and the data parameters are respectively input into the hand-eye matrix to obtain the true value of the hand-eye matrix ⁿ x。

Step S12: three groups of nominal measurement data { of non-parallel mechanical arm rotation axes under the noiseless condition are screened out according to the visual image information and the mechanical arm information ⁿ A _i , ⁿ B _i }。

Where i=1, 2,3 and the rotation angle value of the nominal measurement data when i=3 is constrained to be equal to pi radians.

Specifically, the actual measurement data { is determined according to the visual image information and the mechanical arm information ^a A _i , ^a B _i Homogeneous transformation matrix when i=1 in }Pose matrix->Actual measurement data { ^a A _i , ^a B _i 1.ltoreq.i in }<Homogeneous transformation matrix at 3->Pose matrix- >Actual measurement data { ^a A _i , ^a B _i Homogeneous transformation matrix +.f when i=3 in }>Pose matrix->

Wherein the homogeneous transformation matrixIn order to change the mechanical tool system from the initial state to the first configuration in the visual range of the four-eye camera, the mechanical tool system is relative to the homogeneous transformation matrix of the basic system, the homogeneous transformation matrix is->In order to change the mechanical tool system from the first configuration to the second configuration to the homogeneous transformation matrix of the basic system in the visual range of the four-eye camera, the homogeneous transformation matrix is->In order to be in the visual range of the four-eye camera, the mechanical tool system after the mechanical arm is transformed from the second configuration to the third configuration is in a homogeneous transformation matrix relative to the basic system.

Wherein, the pose matrixIn order to obtain a position and posture matrix of a calibration plate under a camera coordinate system in a calibration plate image shot by a four-eye camera after the mechanical arm is transformed to a first configuration, a position and posture matrix is ∈ ->In order to obtain a position and posture matrix of a calibration plate under a camera coordinate system in a calibration plate image shot by a four-eye camera after the mechanical arm is transformed to a second configuration, and obtain a posture matrixThe system is a position and posture matrix of the calibration plate under a camera coordinate system in the calibration plate image shot by the four-eye camera after the mechanical arm is transformed to the third configuration.

Specifically, according to the mechanical arm information, the homogeneous transformation matrix of the mechanical arm tool system relative to the mechanical arm basic system can be obtained when the mechanical arm is adjusted to a certain configuration. And extracting a position and posture (abbreviated as pose) matrix of the calibration plate under a four-eye camera coordinate system by using a Camera Calibration calibration kit of Matlab R2016 a software based on visual image information.

Calculating homogeneous transformation matrix in two adjacent movements to obtain transition matrixCalculating the pose matrix between two adjacent motions to obtain a relative pose matrix +.>Judging whether j in the transition matrix and the relative pose matrix is equal to a preset value, if not, increasing the value of j (j=j+1), and circularly executing the steps until j is equal to the preset value;

acquiring the acquisition times of data corresponding to the transition matrix and the relative pose matrix, judging whether the acquisition times exceed the preset acquisition times, if not, repeating the steps until the data acquisition times exceed the preset acquisition times, if so, determining the transition matrix data and the relative pose matrix data corresponding to each acquisition time, and screening the transition matrix data and the relative pose matrix data according to preset screening conditions to obtain three groups of data with non-parallel mechanical arm rotation axes.

Specifically, as shown in fig. 3, the data acquisition number is set to i, the pose change number j is initialized, that is, when data screening is performed, the data acquisition number i=1, the pose change number j=1, under the condition that the pose of the machine wall tool system in the basic system is recordedPose of checkerboard calibration plate under four-eye camera +.>Then judging whether the pose change times j is equal to the preset valueSetting a preset value to be 3 in the implementation of the application, namely judging whether the pose change times j is equal to 3, when the pose change times j are not equal to 3, adjusting the mechanical arm to another configuration, and recording a homogeneous transformation matrix ∈>Pose matrix->When the pose change times j is equal to 3, the four-eye camera rotates 178 degrees around the axis through the movement of the mechanical arm, and the homogeneous transformation matrix is recorded>Pose matrix->Tying a mechanical wall tool in the pose of the foundation system>Homogeneous transformation matrix->Bringing into a transition matrix, solving the transition matrix +.>Pose of checkerboard calibration plate under four-eye camera>Pose matrixIs carried into the relative pose matrix, and the relative pose matrix is solved>Adding value based on original pose change times j and data acquisition times i, adding value unit 1, and judging In the embodiment of the present application, the preset collection number is 60, that is, it is determined whether the data collection number is i exceeds 60, if not, the logic is executed circularly, if yes, the random function is used to obtain 60 sets of measurement data sets @, where the number is i exceeds the preset collection number, and if yes, the logic is obtained circularly ^a A _j , ^a B _j ) And randomly screening three groups of data to obtain data of which the rotation axes of the three groups of mechanical arms are not parallel.

Step S13: gauss noise is added to the three groups of sense measurement data to obtain actual measurement data { corresponding to the three groups of sense measurement data under the noise condition respectively ^a A _i , ^a B _i }。

Wherein the upper left mark a is the abbreviation of actual and is used for calibrating the actual measurement data.

Step S14: and respectively extracting measurement data of the three groups of actual measurement data to obtain a rotation axis vector, a rotation angle value and a position column vector corresponding to each group of actual measurement data.

Step S15: and carrying out solution analysis on the rotation axis vector, the rotation angle value and the position column vector corresponding to each group of actual measurement data based on different parameter analysis logics to obtain hand-eye matrix estimated values of different combinations of actual measurement data corresponding to different parameter analysis.

Specifically, the rotation axis vector, the rotation angle value and the position column vector are converted into Euler axis angle parameters, and the Euler axis angle parameters are substituted into a hand-eye calibration analysis solution of the Euler axis angle parameters, so that the first two groups of actual measurement data { are obtained ^a A _i , ^a B _i Hand-eye matrix estimation value corresponding to (i=1, 2)Three sets of actual measurement data { ^a A _i , ^a B _i Hand-eye matrix estimation value corresponding to (i=1, 2, 3)>

The rotation axis vector, the rotation angle value and the position are alignedThe quantity is converted into modified Rodrigas parameters, and the modified Rodrigas parameters are substituted into a hand-eye calibration analysis solution of the modified Rodrigas parameters, so that the first two groups of actual measurement data { are obtained ^a A _i , ^a B _i Hand-eye matrix estimation value corresponding to (i=1, 2)Three sets of actual measurement data { ^a A _i , ^a B _i Hand-eye matrix estimation value corresponding to (i=1, 2, 3)>

Converting the rotation axis vector, the rotation angle value and the position column vector into Euclidean group parameters, substituting the Euclidean group parameters into a hand-eye calibration analysis solution of the Euclidean group parameters to obtain the first two groups of actual measurement data { ^a A _i , ^a B _i Hand-eye matrix estimation value corresponding to (i=1, 2)Three sets of actual measurement data { ^a A _i , ^a B _i Hand-eye matrix estimation value corresponding to (i=1, 2, 3) >

Step S16: and performing precision evaluation analysis on the hand-eye matrix estimated values of different combinations of actual measurement data corresponding to different parameter analyses to obtain an angle error value.

Specifically, each set of actual test data { is judged according to the rotation angle value ^a A _i , ^a B _i In } ^a A _i Is a rotation matrix R of (2) _Ai The corresponding rotation angle value and the actual test data { ^a A _i , ^a B _i In } ^a B _i Is a rotation matrix R of (2) _Bi Angle difference of corresponding rotation angle valueIf the value accords with the preset difference range, rotating the matrix R _Ai Rotation angle value of (2) and rotation matrix R _Bi The rotation angle value of (2) is limited to a range of (0, pi) or (pi, 2 pi).

For the embodiments of the present application, to evaluate the accuracy of various hand-eye calibration analytical solutions, an angular distance error criterion is selected for measuring the accuracy of the solved rotation matrix:

obtaining an angular distance error criterion d _∠ Angular distance error criterion d _∠ Is used for measuring the rotation matrix precision of the hand-eye matrix estimated value corresponding to the first two groups of actual measurement data and the third group of actual measurement data of different parameter analysis,

where logm is the matrix logarithm operator,for the robot rotation data in the hand-eye matrix estimation, ⁿ R _x is true value ⁿ Manipulator rotation data in matrix estimation value corresponding to x, II _F Is the Frobenius norm operator, θ _δ Reading a rotation angle;

substituting the hand-eye matrix estimated values of different combinations of actual measurement data corresponding to different parameter analyses intoTrue value ⁿ Substituting the rotation data of the mechanical arm in x into ⁿ R _x And calculating to obtain angle error values corresponding to different combinations of actual measurement data corresponding to different parameter analyses.

Further, in order to further evaluate the accuracy of the hand-eye matrix solved by the different hand-eye calibration methods, considering that the pose between the mechanical arm tool system and the clamped camera coordinate system is unknown, It is necessary to introduce a rotational error metric According to the rotation error measurement criterion E _r And obtaining the rotation matrix error values given by the hand-eye calibration analysis solutions of the different parameters, and merging the error values into the value angle error values to verify the hand-eye calibration singular phenomenon. .

Step S17: { based on three sets of actual measurement data ^a A _i , ^a B _i Go and true value ⁿ x and angle error values determine three sets of actual measurement data { ^a A _i , ^a B _i And calibrating the singular data group of the singular phenomenon by the hand and eye, and carrying out measurement data deviation correction on the singular data group according to the angle error value.

As shown in fig. 4, the random selection number k of measurement data is set to be 1, then the first selected measurement data is solved based on the hand-eye matrix estimated value and the rotation matrix estimated value error of the hand-eye calibration analysis method of different parameters, the hand-eye matrix estimated value and the rotation matrix estimated value error of the hand-eye calibration analysis method of different parameters are obtained, then whether the random selection number k exceeds the preset selection number is judged, if yes, the data with the error seriously deviated from the true value is selected, the singular phenomenon of the hand-eye calibration analysis method corresponding to the data is verified, if not, the value is increased on the basis of the original random selection number k, the value added value is 1, and the random screening of the measurement data set is circularly executed until the random selection number k exceeds the preset selection number.

In the embodiment of the present application, the preset selection times are exemplified by 160 times, including but not limited to 160 times.

The experimental data obtained by 131 th collection are shown in table 1, table 2 and table 3 in sequence.

Table 1 shows homogeneous coordinate matrix ^a A ₁ And (3) with ^a B ₁ The first 3 rows of elements:

element number	^a A ₁	^a B ₁
			(1,1)	0.9664958772236	0.9700131334560
(1,2)	0.2405337043912	0.2125713887554
			(1,3)	-0.0896061178796	-0.1178470432613
(1,4)	0.1667514779623	0.2751094975113
			(2,1)	-0.2326873859421	-0.1974718492165
(2,2)	0.9684018101380	0.9719460012859
			(2,3)	0.0897469472725	0.1277726079845
(2,4)	0.1343605834941	0.1459244337216
			(3,1)	0.1083618924393	0.1417017631853
(3,2)	-0.065889841198	-0.1006696342834
			(3,3)	0.9919255159001	0.9847772514855
(3,4)	0.3679020880331	0.3071464396775

Table 2 shows homogeneous coordinate matrix ^a A ₂ And (3) with ^a B ₂ The first 3 rows of elements:

element number	^a A ₂	^a B ₂
			(1,1)	-0.2917567339496	-0.3847908803664
(1,2)	0.0023536817001	0.0226563998432
			(1,3)	0.9564896593155	0.9227256720894
(1,4)	-0.4412342700205	-0.4898692143796
			(2,1)	-0.0058037671322	0.0154187445394
(2,2)	-0.9999829196767	-0.9994014145730
			(2,3)	0.0006903925713	0.0309689338904
(2,4)	-0.2079740666131	-1.1573886327017
			(3,1)	0.9564749471272	0.9228749864979
(3,2)	-0.0053498165652	0.0261438347536
			(3,3)	0.2917654108707	0.3842112169116
(3,4)	-0.0151537164186	0.0204890008555

Table 3 shows homogeneous coordinate matrix ^a A ₃ And (3) with ^a B ₃ The first 3 rows of elements:

specifically, as can be seen from fig. 5, 160 sets of constant logarithm values of the rotation error are obtained. Wherein, "MRP", "QA", "EG", "ODT", "DQ" and "EAA" represent errors in rotation matrix estimates given by hand-eye calibration solutions based on modified Rodrigas parameters, quaternion parameters, euclidean group parameters, orthogonal dual tensor parameters, dual quaternion/helix theoretical parameters and Euler axis angle parameters, respectively.

For the rest 159 experiments except the 131 st calibration experiment, the rotation error value corresponding to the existing hand-eye calibration analysis method belongs to the normal error range, and on the contrary, in the 131 st calibration experiment, the rotation error value "MRP", "QA", "EG", "ODT", "DQ" and "EAA" of the hand-eye matrix are all between 0.31 radian and 1.25 radian, and the rotation matrix estimated value given by the existing hand-eye calibration analysis method deviates from the true value seriously, namely the singular phenomenon appears in the hand-eye calibration analysis method, so that the existence of the singular phenomenon shared by the existing hand-eye calibration analysis method is verified. And carrying out measured data deviation correction according to the rotation matrix estimated value and the error value corresponding to the rotation matrix estimated value.

The embodiment of the application provides a method for detecting singular value resolution of robot hand-eye calibration, which is used for acquiring visual image information and mechanical arm information when measuring data deviation correction is carried out on singular data generated by hand-eye calibration, wherein the visual image information is used for representing visual image information shot in a history period by a four-eye camera arranged on a robot mechanical arm, the mechanical arm information is used for representing movement information of the robot mechanical arm in the history period, the measurement data of the mechanical arm is acquired according to the visual image information and the mechanical arm information, and then true value of a hand-eye matrix is set ⁿ x, namely, according to the position column vector of the mechanical arm and the nominal hand-eye calibration analysis value corresponding to the rotation data of the mechanical arm, three groups of nominal measurement data { of non-parallel rotation axes of the mechanical arm under the noiseless condition are screened out according to the visual image information and the mechanical arm information ⁿ A _i , ⁿ B _i Then Gauss noise is added to the three groups of sense measurement data to obtain actual measurement data { corresponding to the three groups of sense measurement data under the noise condition respectively ^a A _i , ^a B _i Because of a certain noise interference in the normal hand-eye calibration analysis process, adding the nominal measurement data with non-parallel rotation axes of the mechanical arm into noise to obtain actual measurement data, improving the authenticity of the measurement data, and respectively extracting the measurement data of three groups of actual measurement data to obtain each group of actual measurement data The method comprises the steps of carrying out solution analysis on a rotation axis vector, a rotation angle value and a position column vector corresponding to data according to different parameter analysis logics mentioned in the background art, obtaining hand-eye matrix estimated values of different combinations of actual measurement data corresponding to different parameter analysis, carrying out precision evaluation analysis on the hand-eye matrix estimated values of different combinations of actual measurement data corresponding to different parameter analysis, obtaining angle error values, and then carrying out { based on three groups of actual measurement data ^a A _i , ^a B _i Go and true value ⁿ x and angle error values determine three sets of actual measurement data { ^a A _i , ^a B _i And the singular data set of the hand-eye calibration singular phenomenon exists in the step, and the measured data deviation correction is carried out on the singular data set according to the angle error value, so that the measured data in the singular data set is corrected in a deviation mode while the hand-eye calibration singular phenomenon is verified, and the safety of a hand-eye calibration algorithm is improved.

The above embodiment describes a method for detecting singular solutions of calibration and analysis of a robot hand and eyes from the viewpoint of a method flow, and the following embodiment describes a device for detecting singular solutions of calibration and analysis of a robot hand and eyes from the viewpoint of a virtual module or a virtual unit, and the following embodiment is described in detail.

The embodiment of the application provides a device 60 for detecting singular values of calibration analysis of a robot hand and a human eye, as shown in fig. 6, fig. 6 is a schematic structural diagram of the device for detecting singular values of calibration analysis of a robot hand and a human eye. The apparatus 60 may specifically include:

the information acquisition module 61 is configured to acquire visual image information and mechanical arm information, where the visual image information is used to represent visual image information captured by a four-eye camera mounted on the mechanical arm of the robot in a history period, and the mechanical arm information is used to represent motion information of the mechanical arm of the robot in the history period;

a truth value setting module 62 for setting truth values of the hand-eye matrix ⁿ x，Wherein, ⁿ t represents a true value ⁿ x mechanical arm position column vector, rot @ ⁿ k, ⁿ θ _x ) Representing true values ⁿ The robot arm of x rotates the data, ⁿ θ _x is Rot% ⁿ k, ⁿ θ _x ) The upper left hand sign n is the abbreviation of nominal for calibrating nominal data;

the data screening module 63 is configured to screen three sets of nominal measurement data { about non-parallel rotation axes of the robot arm under a noiseless condition according to the visual image information and the robot arm information ⁿ A _i , ⁿ B _i -wherein i = 1,2,3 and the rotation angle value of the nominal measurement data when i = 3 is constrained to be equal to pi radians;

A noise mixing module 64 for adding Gauss noise to the three sets of sense measurement data to obtain actual measurement data { corresponding to the three sets of sense measurement data under noise conditions ^a A _i , ^a B _i -wherein the upper left hand mark a is an abbreviation for actual for calibrating the actual measurement data;

the data extraction module 65 is configured to extract measurement data from three sets of actual measurement data, so as to obtain a rotation axis vector, a rotation angle value, and a position column vector corresponding to each set of actual measurement data;

the data analysis module 66 is configured to perform solution analysis on the rotation axis vector, the rotation angle value and the position column vector corresponding to each set of actual measurement data based on different parameter analysis logic, so as to obtain hand-eye matrix estimation values of different combinations of actual measurement data corresponding to different parameter analysis;

the error analysis module 67 is configured to perform accuracy evaluation analysis on hand-eye matrix estimation values of different combinations of actual measurement data corresponding to different parameter analyses, so as to obtain an angle error value;

a singular determination module 68 for { based on three sets of actual measurement data ^a A _i , ^a B _i Go and true value ⁿ x and angle error values determine three sets of actual measurement data { ^a A _i , ^a B _i Hand-eye marks present in }And (3) setting a singular data group of the singular phenomenon, and carrying out measurement data deviation correction on the singular data group according to the angle error value.

For the embodiment of the application, when the measured data deviation correction is performed on the singular data generated by hand eye calibration, visual image information and mechanical arm information are acquired, wherein the visual image information is used for representing visual image information shot in a history period by a four-eye camera mounted on a mechanical arm of a robot, the mechanical arm information is used for representing movement information of the mechanical arm of the robot in the history period, the measured data of the mechanical arm is acquired according to the visual image information and the mechanical arm information, and then a true value of a hand-eye matrix is set ⁿ x, namely, according to the position column vector of the mechanical arm and the nominal hand-eye calibration analysis value corresponding to the rotation data of the mechanical arm, three groups of nominal measurement data { of non-parallel rotation axes of the mechanical arm under the noiseless condition are screened out according to the visual image information and the mechanical arm information ⁿ A _i , ⁿ B _i Then Gauss noise is added to the three groups of sense measurement data to obtain actual measurement data { corresponding to the three groups of sense measurement data under the noise condition respectively ^a A _i , ^a B _i Because of a certain noise interference in the normal hand-eye calibration analysis process, adding noise into nominal measurement data with non-parallel rotating shafts of a mechanical arm to obtain actual measurement data, improving the authenticity of the measurement data, respectively extracting measurement data from three groups of actual measurement data to obtain rotating shaft vectors, rotating angle values and position column vectors corresponding to each group of actual measurement data, solving and analyzing the rotating shaft vectors, the rotating angle values and the position column vectors corresponding to each group of actual measurement data according to different parameter analysis logics mentioned in the background art to obtain hand-eye matrix estimated values of different combinations of actual measurement data corresponding to different parameter analysis, performing precision evaluation analysis on hand-eye matrix estimated values of different combinations of actual measurement data corresponding to different parameter analysis to obtain angle error values, and then carrying out { based on the three groups of actual measurement data ^a A _i , ^a B _i Go and true value ⁿ x and angle error values determine three sets of actual measurement data { ^a A _i , ^a B _i And the singular data set of the hand-eye calibration singular phenomenon exists in the step, and the measured data deviation correction is carried out on the singular data set according to the angle error value, so that the measured data in the singular data set is corrected in a deviation mode while the hand-eye calibration singular phenomenon is verified, and the safety of a hand-eye calibration algorithm is improved.

In one possible implementation manner of the embodiment of the present application, the apparatus 60 further includes: the difference judging module and the angle value limiting module, wherein,

the difference judging module is used for judging each group of actual test data { according to the rotation angle value ^a A _i , ^a B _i In } ^a A _i Is a rotation matrix R of (2) _Ai The corresponding rotation angle value and the actual test data { ^a A _i , ^a B _i In } ^a B _i Is a rotation matrix R of (2) _Bi Whether the rotation angle difference value of the corresponding rotation angle value accords with a preset difference value range or not;

the angle value limiting module is used for limiting the rotation matrix R when the rotation angle difference value accords with a preset difference value range _Ai Rotation angle value of (2) and rotation matrix R _Bi The rotation angle value of (2) is limited to a range of (0, pi) or (pi, 2 pi).

In another possible implementation manner of this embodiment of the present application, when the data analysis module 66 performs solution analysis on the rotation axis vector, the rotation angle value and the position column vector corresponding to each set of actual measurement data based on different parameter analysis logic, the data analysis module is specifically configured to:

Converting the rotation axis vector, the rotation angle value and the position column vector into Euler axis angle parameters, substituting the Euler axis angle parameters into a hand-eye calibration analysis solution of the Euler axis angle parameters to obtain the first two groups of actual measurement data { ^a A _i , ^a B _i Hand-eye matrix estimation value corresponding to (i=1, 2)Three sets of actual measurement data { ^a A _i , ^a B _i Hand-eye matrix estimation value corresponding to (i=1, 2, 3)

Converting the rotation axis vector, the rotation angle value and the position column vector into modified Rodrigas parameters, substituting the modified Rodrigas parameters into a hand-eye calibration analysis solution for modifying the Rodrigas parameters to obtain the first two groups of actual measurement data { ^a A _i , ^a B _i Hand-eye matrix estimation value corresponding to (i=1, 2)Three sets of actual measurement data { ^a A _i , ^a B _i Hand-eye matrix estimation value corresponding to (i=1, 2, 3)>

Converting the rotation axis vector, the rotation angle value and the position column vector into quaternion parameters, and substituting the quaternion parameters into a hand-eye calibration analysis solution of the quaternion parameters to obtain the first two groups of actual measurement data { ^a A _i , ^a B _i Hand-eye matrix estimation value corresponding to (i=1, 2)Three sets of actual measurement data { ^a A _i , ^a B _i Hand-eye matrix estimation value corresponding to (i=1, 2, 3)>

In another possible implementation manner of the embodiment of the present application, the error analysis module 67 performs accuracy evaluation analysis on the hand-eye matrix estimated values of different combinations of actual measurement data corresponding to different parameter analyses, so as to obtain an angle errorThe values, in particular, are: obtaining an angular distance error criterion d _∠ Angular distance error criterion d _∠ Is used for measuring the rotation matrix precision of the hand-eye matrix estimated value corresponding to the first two groups of actual measurement data and the third group of actual measurement data of different parameter analysis,

In another possible implementation manner of this embodiment of the present application, the data filtering module 63 filters three sets of nominal measurement data { for the non-parallel rotation axes of the robot arm under the noiseless condition according to the visual image information and the robot arm information ⁿ A _i , ⁿ B _i When } is used, in particular:

obtain the actual measurement data { ^a A _i , ^a B _i Homogeneous transformation matrix when i=1 in }Pose matrix->Actual measurement data { ^a A _i , ^a B _i 1.ltoreq.i in }<Homogeneous transformation matrix at 3- >Pose matrix->Actual measurement data { ^a A _i , ^a B _i Homogeneous transformation matrix +.f when i=3 in }>Pose matrix->

Wherein the homogeneous transformation matrixIn order to change the mechanical tool system from the initial state to the first configuration in the visual range of the four-eye camera, the mechanical tool system is relative to the homogeneous transformation matrix of the basic system, the homogeneous transformation matrix is->In order to change the mechanical tool system from the first configuration to the second configuration to the homogeneous transformation matrix of the basic system in the visual range of the four-eye camera, the homogeneous transformation matrix is->In order to be in the visual range of the four-eye camera, the mechanical tool system after the mechanical arm is transformed from the second configuration to the third configuration is relative to the homogeneous transformation matrix of the basic system;

wherein, the pose matrixIn order to obtain a position and posture matrix of a calibration plate under a camera coordinate system in a calibration plate image shot by a four-eye camera after the mechanical arm is transformed to a first configuration, a position and posture matrix is ∈ ->In order to obtain a position and posture matrix of a calibration plate under a camera coordinate system in a calibration plate image shot by a four-eye camera after the mechanical arm is transformed to a second configuration, and obtain a posture matrixThe position and posture matrix of the calibration plate under the camera coordinate system in the calibration plate image shot by the four-eye camera after the mechanical arm is transformed to the third configuration;

It will be clearly understood by those skilled in the art that, for convenience and brevity of description, a specific working process of the above-described device 60 for detecting the resolution of the singular value of the calibration of the eyes of the robot hand may refer to a corresponding process in the foregoing method embodiment, which is not described herein again.

An electronic device is provided in an embodiment of the present application, as shown in fig. 7, and fig. 7 is a schematic structural diagram of the electronic device provided in the embodiment of the present application. The electronic device 700 shown in fig. 7 includes: a processor 701 and a memory 703. The processor 701 is coupled to a memory 703, such as via a bus 702. Optionally, the electronic device 700 may also include a transceiver 704. It should be noted that, in practical applications, the transceiver 704 is not limited to one, and the structure of the electronic device 700 is not limited to the embodiments of the present application.

The processor 701 may be a CPU (Central Processing Unit ), general purpose processor, DSP (Digital Signal Processor, data signal processor), ASIC (Application Specific Integrated Circuit ), FPGA (Field Programmable Gate Array, field programmable gate array) or other programmable logic device, transistor logic device, hardware components, or any combination thereof. Which may implement or perform the various exemplary logic blocks, modules, and circuits described in connection with this disclosure. The processor 701 may also be a combination that performs computing functions, such as including one or more microprocessors, a combination of a DSP and a microprocessor, or the like.

Bus 702 may include a path to transfer information between the components. Bus 702 may be a PCI (Peripheral Component Interconnect, peripheral component interconnect Standard) bus or an EISA (Extended Industry Standard Architecture ) bus, or the like. Bus 702 may be divided into an address bus, a data bus, a control bus, and the like. For ease of illustration, only one thick line is shown in fig. 7, but not only one bus or type of bus.

The Memory 703 may be, but is not limited to, ROM (Read Only Memory) or other type of static storage device that can store static information and instructions, RAM (Random Access Memory ) or other type of dynamic storage device that can store information and instructions, EEPROM (Electrically Erasable Programmable Read Only Memory ), CD-ROM (Compact Disc Read Only Memory, compact disc Read Only Memory) or other optical disk storage, optical disk storage (including compact discs, laser discs, optical discs, digital versatile discs, blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium that can be used to carry or store desired program code in the form of instructions or data structures and that can be accessed by a computer.

The memory 703 is used for storing application program codes for executing the present application and is controlled by the processor 701 for execution. The processor 701 is configured to execute application code stored in the memory 703 to implement what is shown in the foregoing method embodiments.

Among them, electronic devices include, but are not limited to: mobile terminals such as mobile phones, notebook computers, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), in-vehicle terminals (e.g., in-vehicle navigation terminals), and the like, and stationary terminals such as digital TVs, desktop computers, and the like. But may also be a server or the like. The electronic device shown in fig. 3 is only an example and should not be construed as limiting the functionality and scope of use of the embodiments herein.

The present application provides a computer readable storage medium having a computer program stored thereon, which when run on a computer, causes the computer to perform the corresponding method embodiments described above.

It should be understood that, although the steps in the flowcharts of the figures are shown in order as indicated by the arrows, these steps are not necessarily performed in order as indicated by the arrows. The steps are not strictly limited in order and may be performed in other orders, unless explicitly stated herein. Moreover, at least some of the steps in the flowcharts of the figures may include a plurality of sub-steps or stages that are not necessarily performed at the same time, but may be performed at different times, the order of their execution not necessarily being sequential, but may be performed in turn or alternately with other steps or at least a portion of the other steps or stages.

The foregoing is only a partial embodiment of the present application and it should be noted that, for a person skilled in the art, several improvements and modifications can be made without departing from the principle of the present application, and these improvements and modifications should also be considered as the protection scope of the present application.

Claims

1. The method for detecting the resolution singularity of the robot hand-eye calibration analysis is characterized by comprising the following steps of:

2. The method for detecting singular resolution of calibration analysis of a robot hand and eye according to claim 1, wherein the extracting of measurement data is performed on three sets of actual measurement data to obtain a rotation axis vector, a rotation angle value and a position column vector corresponding to each set of actual measurement data, and then the method further comprises:

if the rotation angle difference accords with the preset difference range, the rotation matrix R is obtained _Ai Is rotated by (a)Rotation angle value and the rotation matrix R _Bi The rotation angle value of (2) is limited to a range of (0, pi) or (pi, 2 pi).

3. The method for detecting singular resolution of calibration analysis of a robot hand and eye according to claim 2, wherein the solving and analyzing the rotation axis vector, the rotation angle value and the position column vector corresponding to each set of the actual measurement data based on different parameter analysis logic to obtain hand and eye matrix estimated values of different combinations of actual measurement data corresponding to different parameter analysis comprises:

Converting the rotation axis vector, the rotation angle value and the position column vector into Euler axis angle parameters, substituting the Euler axis angle parameters into a hand-eye calibration analysis solution of the Euler axis angle parameters to obtain the first two groups of actual measurement data { ^a A _i , ^a B _i Hand-eye matrix estimation value corresponding to (i=1, 2)Three sets of actual measurement data { ^a A _i , ^a B _i Hand-eye matrix estimation value corresponding to (i=1, 2, 3)Converting the rotation axis vector, the rotation angle value and the position column vector into modified Rodrigas parameters, and substituting the modified Rodrigas parameters into a hand-eye calibration analysis solution for modifying the Rodrigas parameters to obtain the first two groups of actual measurement data { ^a A _i , ^a B _i Hand-eye matrix estimation value corresponding to (i=1, 2)>Three sets of actual measurement data { ^a A _i , ^a B _i Hand-eye matrix estimation value corresponding to (i=1, 2, 3)>

Converting the rotation axis vector, the rotation angle value and the position column vector into quaternion parameters, and substituting the quaternion parameters into a hand-eye calibration analysis solution of the quaternion parameters to obtain the first two groups of actual measurement data { ^a A _i , ^a B _i Hand-eye matrix estimation value corresponding to (i=1, 2)Three sets of actual measurement data { ^a A _i , ^a B _i Hand-eye matrix estimation value corresponding to (i=1, 2, 3)Converting the rotation axis vector, the rotation angle value and the position column vector into Euclidean group parameters, substituting the Euclidean group parameters into a hand-eye calibration analysis solution of the Euclidean group parameters to obtain the first two groups of actual measurement data { ^a A _i , ^a B _i Hand-eye matrix estimation value corresponding to (i=1, 2)>Three sets of actual measurement data { ^a A _i , ^a B _i Hand-eye matrix estimation value corresponding to (i=1, 2, 3)>Converting the rotation axis vector, the rotation angle value and the position column vector into dual quaternion/spiral theoretical parameters, substituting the dual quaternion/spiral theoretical parameters into a hand-eye calibration analysis solution of the dual quaternion/spiral theoretical parameters to obtain the first two groups of actual measurement data { ^a A _i , ^a B _i Hand-eye matrix estimation value corresponding to (i=1, 2)>Three sets of actual measurement data { ^a A _i , ^a B _i Hand-eye matrix estimation value corresponding to (i=1, 2, 3)>

4. The method for detecting singular values of hand-eye calibration analysis of a robot according to claim 3, wherein the performing precision evaluation analysis on hand-eye matrix estimation values of different combinations of actual measurement data corresponding to different parameter analyses to obtain angle error values includes:

Wherein the logm is a matrix logarithm operator, theFor machines in hand-eye matrix estimatesArm rotation data, said ⁿ R _x Is true value ⁿ Manipulator rotation data in matrix estimation value corresponding to x, II _F Is the Frobenius norm operator, the theta _δ Reading a rotation angle;

5. The method for detecting the resolution singularity of the calibration analysis of the robot hand and the eye according to claim 2, wherein three groups of nominal measurement data { of the non-parallel rotation axes of the mechanical arm under the noiseless condition are screened out according to the visual image information and the mechanical arm information ⁿ A _i , ⁿ B _i -comprising:

determining actual measurement data { according to the visual image information and the mechanical arm information ^a A _i , ^a B _i Homogeneous transformation matrix when i=1 in }Pose matrix->Actual measurement data { ^a A _i , ^a B _i 1.ltoreq.i in }<Homogeneous transformation matrix at time 3Pose matrix->Actual measurement data { ^a A _i , ^a B _i Homogeneous transformation matrix +.f when i=3 in }>Pose matrixWherein the homogeneous transformation matrix ∈ ->In order to change the mechanical tool system from the initial state to the first configuration in the visual range of the four-eye camera, the mechanical tool system is relative to the homogeneous transformation matrix of the basic system, and the homogeneous transformation matrix is>In order to change the mechanical tool system from the first configuration to the second configuration to the homogeneous transformation matrix of the basic system in the visual range of the four-eye camera, the homogeneous transformation matrix>In order to be in the visual range of the four-eye camera, the mechanical tool system after the mechanical arm is transformed from the second configuration to the third configuration is a homogeneous transformation matrix relative to the basic system;

wherein the pose matrixIn order to obtain a position and posture matrix of the calibration plate under a camera coordinate system in the calibration plate image shot by the four-eye camera after the mechanical arm is transformed to the first configuration, the posture matrix is>For calibrating the plate in the calibration plate image shot by the four-eye camera after the mechanical arm is converted to the second configuration A position and pose matrix in the camera coordinate system, said pose matrix +.>The position and posture matrix of the calibration plate under the camera coordinate system in the calibration plate image shot by the four-eye camera after the mechanical arm is transformed to the third configuration;

6. The utility model provides a detect robot hand eye and mark and analyze device of solving singularity which characterized in that includes:

7. An electronic device, comprising:

One or more processors;

a memory;

one or more applications, wherein the one or more applications are stored in the memory and configured to be executed by the one or more processors, the one or more applications configured to: a method for detecting the resolution singularities of a robot hand-eye calibration analysis according to any one of claims 1 to 5 is performed.

8. A computer-readable storage medium having stored thereon a computer program, wherein the program when executed by a processor implements a method of detecting robotic hand-eye calibration resolution singularities according to any one of claims 1 to 5.