High-precision coordinate system calibration and pose estimation method for cradle turntable and robot

Abstract

The application discloses a high-precision coordinate system calibration and pose estimation method for a cradle turntable and a robot. First, a workstation composed of an industrial robot, a vision sensor and a cradle turntable is built, the vision sensor is shot by the robot to shoot a calibration board, and a homogeneous transformation matrix of the sensor relative to a robot flange is obtained by combining kinematic identification and hand-eye calibration. Then, the C-axis and the A-axis are rotated respectively, the sensor detects the origin of the calibration board and fits a space circle, the unit normal vectors and the rotation center coordinates of the two axes are obtained, the homogeneous transformation matrix of the turntable relative to the robot base is constructed, and the offset vector between the C-axis and the A-axis is calculated. Finally, the work plane pose estimation is performed in the transformation sequence of "translation to the center of the C-axis, rotation around the C-axis, translation to the center of the A-axis, rotation around the A-axis". The application can inhibit the positioning error of the robot, compensate the offset of the rotation center of the double-axis, and improve the calibration precision and the pose estimation accuracy.

Description

Technical Field

[0001] This invention belongs to the field of calibration technology for industrial robots and automated equipment. Specifically, it relates to a high-precision coordinate system calibration and pose estimation method for cradle turntables and robots. Background Technology

[0002] With the rapid development of intelligent manufacturing, coordinate system calibration technology is increasingly widely used in various automated equipment. In workstations composed of industrial robots, positioners, and vision sensors, high-precision coordinate system calibration between components has become a key link in ensuring collaborative operation. Through accurate coordinate system calibration and the determination of their positional relationships, each component can perform precise motion control and command transmission based on its own and mutual coordinate systems, thereby further improving the workstation's execution efficiency and machining accuracy.

[0003] Currently, coordinate system calibration technologies in industrial robot workstations are mainly divided into the following categories: (1) calibration of industrial robots and turntables, mainly used in various welding workstations with turntables; (2) calibration of robots and vision sensors, mainly used in robot vision systems with eyes outside the hand or eyes on the hand; (3) calibration of vision sensors and turntables, mainly used in various three-dimensional reconstruction systems. With the continuous development of robot offline programming and online programming technologies, the requirements for coordinate system calibration accuracy in robot workstations are getting higher and higher. To achieve high-precision coordinate system calibration, it is necessary to solve the calibration error introduced by the low positioning accuracy of industrial robots, and at the same time, it is also necessary to consider the simplicity and ease of use of the calibration method to form a scientific and reasonable complete set of automated calibration schemes.

[0004] In particular, the cradle turntable, as a precision positioning device with two degrees of freedom—the A-axis (pitch / rotation axis) and the C-axis (slewing axis)—can achieve multi-angle machining of complex curved surfaces. However, compared with single-axis turntables, the coordinate system calibration of cradle turntables faces more severe technical challenges. For single-axis turntables, calibration mainly involves determining the direction and position of their rotation axes. But for cradle turntables, a key geometric feature is that the rotation centers of their A-axis and C-axis usually do not coincide in space, resulting in an inherent inter-axis offset. In practical applications, when the A-axis and C-axis rotate together, if the inter-axis offset is ignored and the normal vectors of the two axes obtained from calibration are simply combined, a systematic, fundamental error will be introduced into the pose calculation, severely restricting the improvement of machining accuracy.

[0005] In summary, the current coordinate system calibration methods applied to cradle dual-drive turntables and industrial robots have the following problems: 1) They fail to effectively solve the calibration error caused by the low positioning accuracy of industrial robots, resulting in low coordinate system calibration accuracy; 2) They use manual teaching for calibration, making the calibration accuracy heavily dependent on the skills and experience of the operators; 3) Existing calibration methods usually only focus on the calibration of the directional vectors of each axis of the dual-axis turntable, while ignoring the key geometric parameter of the offset of the rotation center of the two axes, resulting in the calibration results not being directly and accurately applied to actual pose calculations; 4) Existing calibration methods lack effective accuracy verification and iterative correction mechanisms, failing to guarantee the reliability and consistency of the calibration results. Summary of the Invention

[0006] In view of this, the purpose of the present invention is to provide a high-precision coordinate system calibration and pose estimation method for cradle turntables and robots, so as to solve the calibration deviation introduced by robot positioning error, accurately compensate for the offset of dual-axis rotation center, eliminate systematic pose calculation error, and significantly improve calibration accuracy and pose estimation accuracy.

[0007] To achieve the above objectives, the present invention provides the following technical solution: A high-precision coordinate system calibration and pose estimation method for cradle turntables and robots includes the following steps: S1: Construct a workstation consisting of an industrial robot, a vision sensor, and a cradle turntable, and install the vision sensor at the end of the industrial robot; S2: Set the calibration plate in a fixed position, and use the industrial robot to drive the vision sensor to move to multiple poses, take pictures of the calibration plate to obtain image information, and simultaneously record the angle values ​​of each joint of the industrial robot; S3: Establish and identify the kinematic model of the industrial robot, calculate the robot pose parameters using the identified kinematic model and the joint angle values, and perform hand-eye calibration by combining the homogeneous transformation matrix of the vision sensor relative to the calibration plate to obtain the homogeneous transformation matrix of the vision sensor relative to the flange coordinate system of the industrial robot. S4: Fix the calibration plate to the cradle turntable, rotate the C-axis of the cradle turntable to multiple angles, acquire the position data of the origin of the calibration plate through the vision sensor, and fit a spatial circle to calculate the coordinates of the first circle center. and the unit normal vector of the Z-axis ; S5: Rotate the A-axis of the cradle turntable to multiple angles, acquire the position data of the calibration plate origin through the vision sensor, and fit a spatial circle to calculate the coordinates of the second circle center. and the unit normal vector of the X-axis ; S6: Based on the Z-axis unit normal vector and the unit normal vector of the X-axis Calculate the unit normal vector of the Y-axis Construct the rotation matrix of the cradle turntable relative to the vision sensor, and combine it with the coordinates of the first center of the circle. Or the coordinates of the second center The homogeneous transformation matrix is ​​obtained, and then the homogeneous transformation matrix of the cradle turntable relative to the coordinate system of the industrial robot base when it is at the zero position is calculated in combination with the result of step S3. ; S7: Based on the coordinates of the first center of the circle Second circle center coordinates and the X-axis unit normal vector Calculate the inter-axis offset vector between the C-axis rotation center and the A-axis rotation center of the cradle turntable. ; S8: In subsequent use, based on the inter-axis offset vector Z-axis unit normal vector X-axis unit normal vector and the A-axis rotation angle of the cradle turntable and C-axis rotation angle Calculate the pose of the working plane of the cradle turntable in the coordinate system of the industrial robot base.

[0008] Furthermore, in step S3, the method for identifying the kinematic model of the industrial robot is as follows: A nonlinear least squares algorithm is used to optimize the kinematic parameters by minimizing the Euclidean distance difference between the robot flange pose transformation and the visual sensor pose transformation, expressed as:

[0009] in: This represents the value of the loss function; The kinematic model of the industrial robot is represented, including the kinematic parameters of the industrial robot; and Let represent the homogeneous transformation matrices of the flange coordinate system of the industrial robot relative to the base coordinate system of the industrial robot, obtained during the i-th and j-th movements, respectively. ; This represents the homogeneous transformation of the vision sensor relative to the flange coordinate system of the industrial robot; and These represent the homogeneous transformation matrices of the vision sensor relative to the calibration plate obtained during the i-th and j-th movements of the industrial robot, respectively. This represents the Euclidean distance.

[0010] Furthermore, in step S7, the inter-axis offset vector The calculation method is as follows: calculate the coordinates from the first center of the circle. The perpendicular vector to the A-axis is represented as:

[0011] Where: the axis of the A-axis is the coordinate system passing through the center of the second circle. And the direction vector is the unit normal vector of the X-axis. A straight line.

[0012] Furthermore, in step S8, the method for calculating the pose of the working plane of the cradle turntable in the coordinate system of the industrial robot base is as follows: S8a: Based on the C-axis rotation angle of the cradle turntable Construct a rotation matrix around the Z-axis ; S8b: Based on the inter-axis offset vector Construct the translation transformation matrix The translation transformation matrix Used to represent translation from the rotation center of the C-axis to the rotation center of the A-axis; S8c: Based on the A-axis rotation angle of the cradle turntable Construct a rotation matrix around the X-axis ; S8d: Calculate the homogeneous transformation matrix of the working plane of the cradle turntable in the coordinate system of the industrial robot base, according to the predefined transformation order. :

[0013] in: For the inter-axis offset vector The constructed translation transformation matrix; and These are the rotation matrices around the Z-axis and X-axis, respectively.

[0014] Furthermore, the predefined transformation order is as follows: Firstly, through The coordinate system is translated from the zero position of the cradle turntable to the C-axis rotation center; Then through To achieve rotation about the C-axis; Then through Translate the coordinate system from the rotation center of the C-axis to the rotation center of the A-axis; Finally passed A rotation about axis A is achieved to obtain the pose of the working plane in the coordinate system of the industrial robot base.

[0015] Furthermore, it also includes step S9: fixing the calibration plate on the working plane of the cradle turntable, controlling the A-axis and C-axis of the cradle turntable to rotate to multiple preset angle combinations, detecting the actual position of the origin of the calibration plate through the vision sensor, and comparing it with the theoretical position calculated according to steps S6 to S8 to calculate the position deviation; when the position deviation is greater than a preset threshold, returning to step S4, iteratively correcting the calibration parameters.

[0016] Furthermore, the steps for calculating the positional deviation are as follows: S9a: Fix the calibration plate onto the working plane of the cradle turntable; S9b: Select no fewer than 3 different angle combinations The A-axis and C-axis of the cradle turntable are controlled to rotate to preset angles respectively; and The first The A-axis rotation angle and C-axis rotation angle in the angle combination; S9c: Under each angle combination, the visual sensor captures an image of the calibration board, and the actual position of the calibration board origin relative to the visual sensor is detected and obtained. ; S9d: Execute steps S6 to S8 to calculate the theoretical position of the calibration plate origin relative to the industrial robot base coordinate system for each angle combination. ; S9e: Calculate the position deviation for each angle combination. :

[0017] Then calculate the average position deviation for all angle combinations. and maximum positional deviation .

[0018] Furthermore, when the average position deviation Greater than the preset first threshold or the maximum positional deviation Greater than the preset second threshold At that time, the actual location data Feedback is sent to steps S4 and S5, utilizing the actual location data. After merging with the original calibration data, the spatial circle equation is refitted, and the Z-axis unit normal vector is iteratively corrected. X-axis unit normal vector and the inter-axis offset vector until the average position deviation Less than the preset first threshold and the maximum positional deviation Less than the preset second threshold .

[0019] Furthermore, in step S4, the C-axis of the cradle turntable is rotated by at least 3 angles; in step S5, the A-axis of the cradle turntable is rotated by at least 3 angles.

[0020] Furthermore, in step S2, the industrial robot drives the vision sensor to move at least 4 times.

[0021] The beneficial effects of this invention are as follows: This invention provides a high-precision coordinate system calibration and pose estimation method for cradle turntables and robots, which has the following technical advantages: (1) Suppress robot positioning error and improve calibration accuracy: This invention introduces the industrial robot kinematic model identification step and uses the nonlinear least squares algorithm to optimize and identify the robot kinematic parameters, which effectively suppresses the calibration error caused by the low absolute positioning accuracy of the robot, overcomes the defect of unreliable calibration results caused by robot positioning deviation in traditional methods, and realizes full-process automated calibration without manual teaching, which significantly reduces the dependence on operators and improves calibration efficiency and consistency. (2) Accurately compensate for the offset of the dual-axis rotation center and eliminate systematic errors: This invention fully considers the key geometric feature that the rotation centers of the A-axis and C-axis of the cradle turntable do not coincide, and innovatively proposes a calculation method and compensation strategy for the inter-axis offset vector. By accurately calibrating the offset distance between the two axes and compensating it in the pose estimation transformation chain, the systematic pose calculation error caused by ignoring the inter-axis offset is eliminated. (3) The calibration results can be directly used for actual processing: the high-precision homogeneous transformation matrix of the cradle turntable relative to the robot base when it is at zero position is obtained. The calibration results, along with the calibrated inter-axis offset vector, give the pose estimation model a clear physical meaning and a rigorous mathematical expression. The calibration results can be directly and accurately applied to subsequent offline programming and online control without additional correction.

[0022] Furthermore, this invention also incorporates a closed-loop accuracy verification and iterative correction mechanism. By comparing the actual detected position with the theoretically calculated position of the calibration plate origin, two evaluation indicators, average deviation and maximum deviation, are constructed. When the deviation exceeds the limit, the measured unknown data is automatically merged with the original calibration data, and the spatial circle and calibration parameters are refitted. This effectively compensates for residual errors, ensures the convergence and reliability of the final calibration parameters, and provides a solid guarantee for high-precision machining applications. Attached Figure Description

[0023] To make the objectives, technical solutions, and beneficial effects of this invention clearer, the following figures are provided for illustration: Figure 1 This is a flowchart of the high-precision coordinate system calibration and pose estimation method for cradle turntables and robots according to the present invention. Figure 2 This is a schematic diagram of the structure of an industrial robot workstation.

[0024] Explanation of reference numerals in the attached figures: 1-Industrial robot; 2-Vision sensor; 3-Cradle turntable. Detailed Implementation

[0025] The present invention will be further described below with reference to the accompanying drawings and specific embodiments, so that those skilled in the art can better understand and implement the present invention. However, the embodiments described are not intended to limit the present invention.

[0026] like Figure 1 As shown, this embodiment describes a high-precision coordinate system calibration and pose estimation method for a cradle turntable and a robot, which includes the following steps.

[0027] S1: Construct a workstation consisting of an industrial robot 1, a vision sensor 2, and a cradle turntable 3, such as... Figure 2 As shown, the vision sensor 2 is mounted on the end of the industrial robot 1.

[0028] S2: Set the calibration plate in a fixed position; the industrial robot 1 drives the vision sensor 2 to move multiple times, capturing image information on the calibration plate and simultaneously acquiring the angle values ​​of each joint of the industrial robot 1. Establish a calibration plate coordinate system and calculate the homogeneous transformation matrix of the vision sensor 2 relative to the calibration plate based on the image information captured by the vision sensor 2. In this embodiment, the industrial robot 1 drives the vision sensor 2 to move at least 4 times, such as 4, 5, or 6 times, etc., which will not be elaborated further.

[0029] S3: Establish the kinematic model of the industrial robot 1, identify the kinematic model of the industrial robot 1, and obtain the identified kinematic model of the industrial robot 1; input the joint angle values ​​of the industrial robot 1 obtained in step 2 into the identified kinematic model of the industrial robot 1, and calculate the pose parameters of the industrial robot 1. Based on the pose parameters of the industrial robot 1 and the homogeneous transformation matrix of the vision sensor 2 relative to the calibration plate obtained in step S2, perform hand-eye calibration to obtain the homogeneous transformation matrix of the vision sensor 2 relative to the flange coordinate system of the industrial robot 1.

[0030] In this embodiment, the method for identifying the kinematic model of the industrial robot 1 is as follows: a nonlinear least squares algorithm is used to optimize the kinematic parameters by minimizing the Euclidean distance difference between the robot flange pose transformation and the vision sensor 2 pose transformation, which is expressed as follows:

[0031] in: This represents the value of the loss function; The kinematic model of the industrial robot 1 is represented, including the kinematic parameters of the industrial robot 1; and Let represent the homogeneous transformation matrices of the flange coordinate system of the industrial robot 1 relative to the base coordinate system of the industrial robot 1, obtained during the i-th and j-th movements, respectively. ; This represents the homogeneous transformation of the vision sensor 2 relative to the flange coordinate system of the industrial robot 1; and These represent the homogeneous transformation matrices of the vision sensor 2 relative to the calibration plate obtained during the i-th and j-th movements of the industrial robot 1, respectively. This represents the Euclidean distance.

[0032] Specifically, the joint angle values ​​of the industrial robot 1 in step S2 are sequentially input into the kinematic model of the industrial robot 1 after identification. Then, the pose parameters of the industrial robot 1 are calculated using the forward kinematics formula. Based on the pose parameters of the industrial robot 1 and the inverse matrix of the homogeneous transformation matrix of the vision sensor 2 relative to the calibration plate in step 2, hand-eye calibration is performed to obtain the homogeneous transformation matrix of the vision sensor 2 relative to the flange coordinate system of the industrial robot 1.

[0033] S4: Rotate the A-axis of the cradle turntable 3 to 0 degrees, keeping the working plane of the cradle turntable 3 horizontal; fix the calibration plate on the cradle turntable 3, rotate the C-axis of the cradle turntable 3 to multiple angles, and simultaneously capture image information on the calibration plate through the vision sensor 2 to obtain the position data of the calibration plate origin relative to the vision sensor 2; obtain the coordinates of the first circle center by fitting the equation of a spatial circle based on the coordinates of multiple calibration plate origins. And the equation of the plane containing the first circle; calculate the normal vector of the plane and normalize it to obtain the unit normal vector of the Z-axis of the cradle turntable 3. In this embodiment, the C-axis of the cradle turntable 3 is rotated to at least three angles, such as three, four, or five angles, which will not be elaborated further.

[0034] S5: Fix the calibration plate onto the cradle turntable 3, rotate the A-axis of the cradle turntable 3 to multiple angles, and simultaneously capture image information on the calibration plate using the vision sensor 2 to obtain the position data of the calibration plate origin relative to the vision sensor 2; fit the equation of a spatial circle based on the coordinates of the multiple calibration plate origins to obtain the coordinates of the second circle center. And the equation of the plane containing the second circle; calculate the normal vector of the plane and normalize it to obtain the unit normal vector of the X-axis of the cradle turntable 3. In this embodiment, the A-axis of the cradle turntable 3 is rotated to at least three angles, such as three, four, or five angles, which will not be elaborated further.

[0035] S6: The Z-axis unit normal vector calculated based on steps S4 and S5 and the X-axis unit normal vector ,according to Calculate the unit normal vector of the Y-axis This allows for the construction of a rotation matrix of the cradle turntable 3 relative to the visual sensor 2; combined with the coordinates of the first center circle. Or the coordinates of the second center of the circle Obtain the homogeneous transformation matrix of the cradle turntable 3 relative to the vision sensor 2; based on the homogeneous transformation matrix of the vision sensor 2 relative to the flange coordinate system of the industrial robot 1 obtained in step 3, calculate the homogeneous transformation matrix of the cradle turntable 3 relative to the base coordinate system of the industrial robot 1. ,in: It is the homogeneous transformation matrix of the cradle turntable 3 relative to the coordinate system of the industrial robot 1 base when it is at the zero position.

[0036] S7: Based on the coordinates of the first center obtained in step S4 The coordinates of the second center of the circle obtained in step S5 and the X-axis unit normal vector Calculate the inter-axis offset vector between the C-axis rotation center and the A-axis rotation center of the cradle turntable 3. .

[0037] In this embodiment, the inter-axis offset vector The calculation method is as follows: calculate the coordinates from the first center of the circle. The perpendicular vector to the A-axis is expressed as:

[0038] Where: the axis of the A-axis is the coordinate system passing through the center of the second circle. And the direction vector is the unit normal vector of the X-axis. A straight line.

[0039] Specifically, inter-axis offset vector Rotation center from the C-axis The perpendicular vector to the A-axis, whose direction is perpendicular to the A-axis (i.e., the X-axis direction), lies within the unit normal vector of the Z-axis. and the Y-axis unit normal vector Within the plane of Zhang Cheng; the inter-axis offset vector The modulus is the coordinate of the first center of the circle. The perpendicular distance to the axis of the A-axis.

[0040] S8: In subsequent use, based on the inter-axis offset vector obtained in step S7 The Z-axis unit normal vector obtained in step S4 The X-axis unit normal vector obtained in step S5 and the A-axis rotation angle of the cradle turntable 3 and C-axis rotation angle Calculate the pose of the working plane of the cradle turntable 3 in the coordinate system of the base of the industrial robot 1.

[0041] In this embodiment, the method for calculating the pose of the working plane of the cradle turntable 3 in the coordinate system of the industrial robot 1 base is as follows: S8a: Based on the C-axis rotation angle of the cradle turntable 3 Construct a rotation matrix around the Z-axis ; S8b: Based on the inter-axis offset vector Construct the translation transformation matrix The translation transformation matrix Used to represent translation from the rotation center of the C-axis to the rotation center of the A-axis; S8c: Based on the A-axis rotation angle of the cradle turntable 3 Construct a rotation matrix around the X-axis ; S8d: Calculate the homogeneous transformation matrix of the working plane of the cradle turntable 3 in the coordinate system of the industrial robot 1 base, according to the predefined transformation order. :

[0042] in: For the inter-axis offset vector The constructed translation transformation matrix; and These are the rotation matrices around the Z-axis and X-axis, respectively.

[0043] Specifically, in step S8d, the physical meaning of the predefined transformation order is as follows: Firstly, through The coordinate system is translated from the zero position of the cradle turntable 3 to the C-axis rotation center; Then through To achieve rotation about the C-axis; Then through Translate the coordinate system from the rotation center of the C-axis to the rotation center of the A-axis; Finally passed A rotation about axis A is achieved to obtain the pose of the working plane in the coordinate system of the industrial robot 1 base.

[0044] S9: Fix the calibration plate on the working plane of the cradle turntable 3, rotate the A-axis and C-axis of the cradle turntable 3 to several preset angle combinations, detect the actual position of the origin of the calibration plate through the vision sensor 2, and compare it with the theoretical position calculated according to steps S6 to S8 to calculate the position deviation; when the position deviation is greater than the preset threshold, return to step S4 and iteratively correct the calibration parameters.

[0045] In this embodiment, the method for calculating the positional deviation is as follows: S9a: Fix the calibration plate onto the working plane of the cradle turntable 3; S9b: Select no fewer than 3 different angle combinations The A-axis and C-axis of the cradle turntable 3 are respectively controlled to rotate to a preset angle; and The first The A-axis rotation angle and C-axis rotation angle in the angle combination; S9c: Under each angle combination, the visual sensor 2 captures an image of the calibration board, and detects and obtains the actual position of the calibration board origin relative to the visual sensor 2. ; S9d: Execute steps S6 to S8 to calculate the theoretical position of the calibration plate origin relative to the coordinate system of the industrial robot 1 base under each angle combination. ; S9e: Calculate the position deviation for each angle combination. :

[0046] Then calculate the average position deviation for all angle combinations. and maximum positional deviation .

[0047] Specifically, when the average position deviation Greater than the preset first threshold or the maximum positional deviation Greater than the preset second threshold At that time, the actual location data Feedback is sent to steps S4 and S5, utilizing the actual location data. After merging with the original calibration data, the spatial circle equation is refitted, and the Z-axis unit normal vector is iteratively corrected. X-axis unit normal vector and the inter-axis offset vector until the average position deviation Less than the preset first threshold and the maximum positional deviation Less than the preset second threshold .

[0048] In summary, this embodiment introduces a kinematic model identification step for the industrial robot 1 and uses a nonlinear least squares algorithm to optimize and identify the robot's kinematic parameters. This effectively suppresses the calibration error caused by the robot's low absolute positioning accuracy, overcomes the defect of unreliable calibration results caused by robot positioning deviation in traditional methods, and achieves fully automated calibration without the need for manual teaching. This significantly reduces reliance on operators and improves calibration efficiency and consistency.

[0049] This embodiment fully considers the key geometric feature that the rotation centers of the A-axis and C-axis of the cradle turntable 3 do not coincide. It innovatively proposes a calculation method and compensation strategy for the inter-axis offset vector. By accurately calibrating the offset distance between the two axes and compensating for it in the pose estimation transformation chain, it eliminates the systematic pose calculation error caused by ignoring the inter-axis offset. It establishes a pose estimation model with clear physical meaning and rigorous mathematical expression, so that the calibration results can be directly and accurately applied to the actual machining process.

[0050] This embodiment also incorporates an accuracy verification and iterative correction mechanism. By comparing the actual detection position of the calibration board origin with the theoretically calculated position, the calibration accuracy is evaluated, and the calibration parameters are automatically fed back for iterative correction when the deviation exceeds the limit. This ensures the reliability and consistency of the calibration results and provides a solid guarantee for high-precision machining applications.

[0051] The above-described embodiments are merely preferred embodiments provided to fully illustrate the present invention, and the scope of protection of the present invention is not limited thereto. Equivalent substitutions or modifications made by those skilled in the art based on the present invention are all within the scope of protection of the present invention. The scope of protection of the present invention is defined by the claims.

Claims

1. A high-precision coordinate system calibration and pose estimation method for cradle turntables and robots, characterized in that: Includes the following steps: S1: Build a workstation consisting of an industrial robot (1), a vision sensor (2) and a cradle turntable (3), and install the vision sensor (2) at the end of the industrial robot (1); S2: Set the calibration plate in a fixed position, and drive the vision sensor (2) to move to multiple poses through the industrial robot (1), take pictures of the calibration plate to obtain image information, and simultaneously record the angle values ​​of each joint of the industrial robot (1); S3: Establish and identify the kinematic model of the industrial robot (1), calculate the robot pose parameters using the identified kinematic model and the joint angle values, and perform hand-eye calibration by combining the homogeneous transformation matrix of the vision sensor (2) relative to the calibration plate to obtain the homogeneous transformation matrix of the vision sensor (2) relative to the flange coordinate system of the industrial robot (1). S4: Fix the calibration plate onto the cradle turntable (3), rotate the C-axis of the cradle turntable (3) to multiple angles, obtain the position data of the calibration plate origin through the vision sensor (2), and fit a spatial circle to calculate the coordinates of the first circle center. and the unit normal vector of the Z-axis ; S5: Rotate the A-axis of the cradle turntable (3) to multiple angles, obtain the position data of the calibration plate origin through the vision sensor (2), and fit a spatial circle to calculate the coordinates of the second circle center. and the unit normal vector of the X-axis ; S6: Based on the Z-axis unit normal vector and the unit normal vector of the X-axis Calculate the unit normal vector of the Y-axis Construct the rotation matrix of the cradle turntable (3) relative to the vision sensor (2), and combine it with the coordinates of the first center of the circle. Or the coordinates of the second center The homogeneous transformation matrix is ​​obtained, and then the homogeneous transformation matrix of the cradle turntable (3) relative to the base coordinate system of the industrial robot (1) at the zero position is calculated by combining the result of step S3. ; S7: Based on the coordinates of the first center of the circle Second circle center coordinates and the X-axis unit normal vector Calculate the inter-axis offset vector between the C-axis rotation center and the A-axis rotation center of the cradle turntable (3). ; S8: In subsequent use, based on the inter-axis offset vector Z-axis unit normal vector X-axis unit normal vector and the A-axis rotation angle of the cradle turntable (3) and C-axis rotation angle Calculate the pose of the working plane of the cradle turntable (3) in the coordinate system of the industrial robot (1) base.

2. The high-precision coordinate system calibration and pose estimation method for cradle turntables and robots according to claim 1, characterized in that: In step S3, the method for identifying the kinematic model of the industrial robot (1) is as follows: a nonlinear least squares algorithm is used to optimize the kinematic parameters by minimizing the Euclidean distance difference between the robot flange pose transformation and the visual sensor (2) pose transformation, which is expressed as: in: This represents the value of the loss function; The kinematic model of the industrial robot (1) is represented, including the kinematic parameters of the industrial robot (1); and Let represent the homogeneous transformation matrices of the flange coordinate system of the industrial robot (1) relative to the base coordinate system of the industrial robot (1) obtained during the i-th and j-th movements, respectively. ; This represents the homogeneous transformation of the vision sensor (2) relative to the flange coordinate system of the industrial robot (1); and Let represent the homogeneous transformation matrices of the vision sensor (2) relative to the calibration plate obtained by the industrial robot (1) during the i-th and j-th movements, respectively; This represents the Euclidean distance.

3. The high-precision coordinate system calibration and pose estimation method for cradle turntables and robots according to claim 1, characterized in that: In step S7, the inter-axis offset vector The calculation method is as follows: calculate the coordinates from the first center of the circle. The perpendicular vector to the A-axis is represented as: Where: the axis of the A-axis is the coordinate system passing through the center of the second circle. And the direction vector is the unit normal vector of the X-axis. A straight line.

4. The high-precision coordinate system calibration and pose estimation method for cradle turntables and robots according to claim 1, characterized in that: In step S8, the method for calculating the pose of the working plane of the cradle turntable (3) in the coordinate system of the industrial robot (1) is as follows: S8a: According to the C-axis rotation angle of the cradle turntable (3) Construct a rotation matrix around the Z-axis ; S8b: Based on the inter-axis offset vector Construct the translation transformation matrix The translation transformation matrix Used to represent translation from the rotation center of the C-axis to the rotation center of the A-axis; S8c: Based on the A-axis rotation angle of the cradle turntable (3) Construct a rotation matrix around the X-axis ; S8d: Calculate the homogeneous transformation matrix of the working plane of the cradle turntable (3) in the coordinate system of the industrial robot (1) according to the predefined transformation order. : in: For the inter-axis offset vector The constructed translation transformation matrix; and These are the rotation matrices around the Z-axis and X-axis, respectively.

5. The high-precision coordinate system calibration and pose estimation method for cradle turntables and robots according to claim 4, characterized in that: The predefined transformation sequence is as follows: Firstly, through The coordinate system is translated from the zero position of the cradle turntable (3) to the C-axis rotation center; Then through To achieve rotation about the C-axis; Then through Translate the coordinate system from the rotation center of the C-axis to the rotation center of the A-axis; Finally passed Achieving rotation around axis A, the pose of the working plane in the coordinate system of the industrial robot (1) base is obtained.

6. The high-precision coordinate system calibration and pose estimation method for cradle turntables and robots according to claim 1, characterized in that: The process also includes step S9: fixing the calibration plate on the working plane of the cradle turntable (3), controlling the A-axis and C-axis of the cradle turntable (3) to rotate to multiple preset angle combinations, detecting the actual position of the calibration plate origin through the vision sensor (2), and comparing it with the theoretical position calculated according to steps S6 to S8 to calculate the position deviation; when the position deviation is greater than a preset threshold, returning to step S4, iteratively correcting the calibration parameters.

7. The high-precision coordinate system calibration and pose estimation method for cradle turntables and robots according to claim 6, characterized in that: The steps for calculating positional deviation are as follows: S9a: Fix the calibration plate on the working plane of the cradle turntable (3); S9b: Select no fewer than 3 different angle combinations The A-axis and C-axis of the cradle turntable (3) are respectively controlled to rotate to a preset angle; and The first The A-axis rotation angle and C-axis rotation angle in the angle combination; S9c: Under each angle combination, the calibration plate image is captured by the vision sensor (2), and the actual position of the calibration plate origin relative to the vision sensor (2) is detected and obtained. ; S9d: Execute steps S6 to S8 to calculate the theoretical position of the calibration plate origin relative to the coordinate system of the industrial robot (1) base under each angle combination. ; S9e: Calculate the position deviation for each angle combination. : Then calculate the average position deviation for all angle combinations. and maximum positional deviation .

8. The high-precision coordinate system calibration and pose estimation method for cradle turntables and robots according to claim 7, characterized in that: When the average position deviation Greater than the preset first threshold or the maximum positional deviation Greater than the preset second threshold At that time, the actual location data Feedback is sent to steps S4 and S5, utilizing the actual location data. After merging with the original calibration data, the spatial circle equation is refitted, and the Z-axis unit normal vector is iteratively corrected. X-axis unit normal vector and the inter-axis offset vector until the average position deviation Less than the preset first threshold and the maximum positional deviation Less than the preset second threshold .

9. The high-precision coordinate system calibration and pose estimation method for cradle turntables and robots according to claim 1, characterized in that: In step S4, the C-axis of the cradle turntable (3) is rotated by at least 3 angles; in step S5, the A-axis of the cradle turntable (3) is rotated by at least 3 angles.

10. The high-precision coordinate system calibration and pose estimation method for cradle turntables and robots according to claim 1, characterized in that: In step S2, the industrial robot (1) drives the vision sensor (2) to move at least 4 times.

Images