CN-122008233-A - Terminal structure for repairing mechanical arm cracks and high-precision calibration method

CN122008233ACN 122008233 ACN122008233 ACN 122008233ACN-122008233-A

Abstract

The invention discloses a tail end structure for repairing a crack of a mechanical arm and a high-precision calibration method, which belong to the technical field of industrial robots and machine vision, wherein a plurality of groups of original pose data for hand-eye calibration are obtained, the plurality of groups of original pose data are preprocessed, effective relative motion pairs are constructed and screened, an analytic initial value of hand-eye transformation is solved based on the obtained effective relative motion pairs, the analytic initial value is taken as a starting point, nonlinear optimization is performed to obtain a final high-precision calibration result, global reasonable initial solution is obtained, optimization is performed, the optimization process is stable and reliable, the space aggregation error of a calibration plate under a base coordinate system is directly minimized, the influence of noise on the solution is effectively suppressed, the noise influence can be remarkably reduced, and the calibration result with higher precision than that of a single analytic method is obtained.

Inventors

YANG XU
XU WEI
XU KUN
WU DAI
WANG HAINIAN

Assignees

长安大学

Dates

Publication Date: 20260512
Application Date: 20260325

Claims (10)

1. The calibration method for the end structure for repairing the mechanical arm crack is characterized by comprising the following steps of: S1, acquiring a plurality of groups of original pose data of hand-eye calibration; S2, preprocessing a plurality of groups of original pose data, and constructing and screening effective relative motion pairs; s3, solving an analysis initial value of hand-eye transformation based on obtaining an effective relative motion pair; and S4, taking the analysis initial value as a starting point, and performing nonlinear optimization to obtain a final high-precision calibration result.
2. The method for calibrating the tail end structure for repairing the cracks of the mechanical arm according to claim 1, wherein the method for acquiring the plurality of groups of original pose data for calibrating the eyes and hands specifically comprises the steps of calibrating a tool center point by a four-point method, controlling the mechanical arm to move to different poses, synchronously executing two operations in each pose, reading a pose matrix of a wrist joint coordinate system relative to a robot base coordinate system from a robot controller, and calculating a pose matrix of a current calibration plate coordinate system relative to a camera coordinate system.
3. The calibration method for the tail end structure for repairing the mechanical arm crack, which is disclosed in claim 2, is characterized by coinciding the origin of the current tool coordinate system with the geometric center of the tail end flange, enabling the gesture to be consistent, selecting a fixed reference point in the working space of the mechanical arm, controlling the mechanical arm to enable the tip of a clamping jaw to sequentially touch the fixed point in at least four different gestures, and recording a homogeneous transformation matrix of the current wrist joint relative to the coordinate system of the base by a controller during each touch to form four groups of and more reference gestures.
4. The method for calibrating the tail end structure for repairing the mechanical arm crack according to claim 1, wherein the preprocessing of the plurality of groups of original pose data and the construction and screening of the effective relative motion pairs are carried out specifically, wherein the collected plurality of groups of original pose data are combined in pairs, the relative motion of the wrist joint of the mechanical arm and the relative motion of a calibration plate are calculated according to an AX=XB model, and the rotation angle of each motion pair is calculated, and only the motion pairs with the rotation angle larger than a preset threshold value are reserved.
5. The calibration method for the tail end structure for repairing the mechanical arm crack according to claim 1, wherein the method for solving the analytic initial value of the hand-eye transformation based on the obtained effective relative motion pair specifically comprises the steps of adopting a danilitis quaternion method, converting a rotation part of the motion pair into a quaternion, constructing an equation set, solving a rotation matrix initial value through singular value decomposition, and solving an initial value of a translation vector through constructing a linear least square equation based on the rotation.
6. The calibration method for the tail end structure for repairing the mechanical arm crack, which is characterized by taking an analysis initial value as a starting point, carrying out nonlinear optimization to obtain a final high-precision calibration result, specifically comprising the steps of carrying out fine solution by adopting Nelder-Mead nonlinear optimization based on geometric consistency, directly minimizing the sum of position variance and attitude angle variance of a predicted pose of a calibration plate under a robot base coordinate system at all sampling moments, taking a parameter vector corresponding to the analysis initial solution as the initial value, iteratively minimizing the target function by utilizing a Nelder-Mead algorithm until convergence, and finally converting the optimized parameter vector back to a homogeneous transformation matrix, thereby obtaining a high-precision final hand-eye external reference matrix.
7. The end structure for repairing the mechanical arm crack according to the method of claim 1, which is characterized by comprising a mechanical arm, an end structure, an RGB-D depth camera, an asymmetric circular array calibration plate and a control unit, wherein the end structure comprises a flange plate connector, an adapter flange plate connector, a crack pouring glue injector connector, a Gemini2 camera and a crack pouring glue injector; The Gemini2 camera is rigidly arranged at the flange at the tail end of the mechanical arm through the flange plate connecting piece and the adapter flange plate connecting piece, so that a fixed eye-on-hand relationship is formed between the camera optical center of the Gemini2 camera and the wrist joint of the mechanical arm; The crack pouring glue injector is rigidly connected with the tail end of the mechanical arm through the crack pouring glue injector connecting piece and is used for executing crack pouring glue injection operation; the asymmetric circular array calibration plate is fixedly arranged in the working space of the mechanical arm and is used for providing unique corner ordering and a calibration plate coordinate system; The control unit comprises a data processing system and a robot controller, wherein the robot controller is connected with the mechanical arm, the tail end structure and the RGB-D camera and is used for controlling the mechanical arm to move and collecting pose and visual data, and the data processing system is used for calibrating the tail end structure for repairing the cracks of the mechanical arm.
8. The end structure of mechanical arm crack repair of claim 7, wherein the data processing system comprises: The tool center point calibration module is used for calibrating a tool center point at the tail end of the mechanical arm through a four-point method to obtain the pose of the tool coordinate system relative to the wrist joint coordinate system; the calibration plate pose solving module is used for detecting the characteristics of the calibration plate and solving a pose matrix of the calibration plate under a camera coordinate system based on the image and the depth data acquired by the RGB-D camera; The calibration data acquisition and pairing module is used for synchronously acquiring a pose matrix from the machine base to the wrist joint and a pose matrix from the camera to the calibration plate under the sampling poses of a plurality of groups of mechanical arms, and constructing a relative motion pair; the hand-eye analysis solving module is used for constructing a hand-eye calibration model based on the relative motion pair, and obtaining an initial hand-eye transformation matrix of the camera relative to the wrist joint by adopting singular value decomposition, namely solving an analysis initial value of hand-eye transformation; And the hand-eye nonlinear optimization module is used for parameterizing the initial hand-eye transformation matrix into a pose vector, constructing a position error and pose error combined objective function, performing global refinement on hand-eye transformation by using a derivative-free optimization algorithm, and outputting a final hand-eye calibration result.
9. The end structure of claim 8, wherein the asymmetric circular array calibration plate is provided with an array of dots with unequal row-column spacing or unique origin marks.
10. The tail end structure of the mechanical arm crack repairing according to claim 8, wherein the calibration plate pose solving module is used for carrying out graying and self-adaptive threshold binarization on RGB images to obtain a binary image with enhanced contrast, extracting circle center candidate points by adopting a circular feature detection algorithm, and carrying out unique sorting on the round points by utilizing asymmetric geometric constraints; The calibration plate pose solving module further comprises a three-dimensional fine solving unit based on a depth map, wherein the three-dimensional fine solving unit is used for extracting depth data of the calibration plate region in a corresponding depth map according to an external polygonal region of the calibration plate in an image and generating local point clouds, performing ICP registration on the local point clouds and a three-dimensional model of the calibration plate by taking poses obtained through a PnP or IPPE algorithm as initial values, and taking rigid transformation output by ICP as fine pose estimation of the calibration plate under a camera coordinate system when the internal point proportion and mean square error of ICP registration meet preset thresholds.

Description

Terminal structure for repairing mechanical arm cracks and high-precision calibration method Technical Field The invention belongs to the technical field of industrial robots and machine vision, and particularly relates to a tail end structure for repairing a crack of a mechanical arm and a high-precision calibration method. Background As infrastructure ages, automated crack detection and repair by robots is an important development direction. The high-precision terminal crack pouring device is a core execution unit for realizing reliable repair operation. Such devices typically integrate visual sensors to form an eye-outside-hand or eye-on-hand sensing system to enable real-time locating of cracks and guiding grouting. In vision-guided robotic tasks, accurate hand-eye calibration is a cornerstone to ensure accuracy of system operation. In a robot grabbing and operating system based on visual guidance, in order to seamlessly connect visual perception to robot control, two types of core space transformation relations must be accurately established, namely, tool center point calibration, position of an end tool relative to a robot flange, and hand-eye calibration, and rigid transformation between a camera coordinate system and a robot coordinate system (an end or a base) are established. The traditional hand-eye calibration method is mostly dependent on an analysis algorithm, and can quickly obtain solutions, but is extremely sensitive to noise of acquired data, and particularly when the robot is limited in motion range or small in relative motion, an equation is easy to be in a pathological state, so that a calibration result is unstable and the precision is limited. Although the prior calibration technology can realize basic functions, the prior hand-eye calibration method has obvious limitation of being extremely sensitive to the geometric configuration of the sampling motion of the robot. In an actual narrow or limited working space (such as a bridge bottom surface and a tunnel side wall), a robot cannot make ideal movements in a large range and multiple angles, so that a calibration equation is in a pathological state, and a resolving result is unstable and has large dispersion. This instability is amplified by the subsequent visual positioning and robot kinematic chain, eventually causing the crack pouring trajectory to shift, affecting the quality of repair. More importantly, the main stream method only optimizes the residual error of the motion equation in the middle process, but does not directly minimize the overall re-projection consistency of the calibration plate under the robot base, so that the requirement of high-precision operation is difficult to meet in terms of precision and robustness. Disclosure of Invention The invention aims to provide a tail end structure for repairing a crack of a mechanical arm and a high-precision calibration method, so as to solve the problems of noise sensitivity and unstable pose resolving in the prior art. In order to achieve the above purpose, the technical scheme adopted by the invention is as follows: a calibration method for an end structure for repairing a crack of a mechanical arm comprises the following steps: S1, acquiring a plurality of groups of original pose data of hand-eye calibration; S2, preprocessing a plurality of groups of original pose data, and constructing and screening effective relative motion pairs; s3, solving an analysis initial value of hand-eye transformation based on obtaining an effective relative motion pair; and S4, taking the analysis initial value as a starting point, and performing nonlinear optimization to obtain a final high-precision calibration result. Preferably, the method for acquiring the multiple groups of original pose data of hand-eye calibration specifically comprises the steps of calibrating a tool center point by adopting a four-point method, controlling a mechanical arm to move to different poses, synchronously executing two operations at each pose, reading a pose matrix of a wrist joint coordinate system relative to a robot base coordinate system from a robot controller, and calculating a pose matrix of a current calibration plate coordinate system relative to a camera coordinate system. The method comprises the steps of enabling an origin of a coordinate system of a current tool to coincide with a geometric center of a flange at the tail end, enabling the gestures to be consistent, selecting a fixed reference point in a working space of a mechanical arm, controlling the mechanical arm to sequentially touch the fixed point by the tip of a clamping jaw under at least four different gestures, and recording a homogeneous transformation matrix of the current wrist joint relative to a coordinate system of a base by a controller when the mechanical arm touches the fixed point each time, so that four groups of reference gestures and more than four groups of reference gestures are formed. Preferably, the preprocessing