Research On Calibration Of Geometric Parameters Of Industrial Robot Based On Vision Detection

Due to the fast,efficient and high precision of industrial robots suits the needs of mod-ern industrial and manufacturin,it is widely used in modern production.However,the ac-curacy of the end positioning of industrial robots that are not calibrated by system precision calibration is very low.In order to improve the end positioning accuracy of industrial robots,in this paper,a six-degree-of-freedom industrial robot independently researched and devel-oped by the enterprise is taken as the research object,and proposed two kinds of geometric parameters calibration and error correction methods for industrial robot based on visual in-spection,and make a study of the improvement of the positioning accuracy of the robot end.This paper mainly studies the following aspects:First,make a study of the robot kinematics.According to the body structure and pa-rameters size of the experimental robot,established the robot joint linkage system and D-H parameter model.Make a intensive deduce of the positive and inverse mathematical models of robot kinematics and determined the key parameters affecting the robot's end position and the mechanism of its influence,provided theoretical support for establishing the robot end position error constraint equation.The principle and experimental scheme of the two parameters calibration methods used in this paper are preliminarily explored,and deter-mined the program of parameter calibration in this paper.Second,make a study of the imaging model of the industrial camera and the calibration of the internal and external parameters of the camera.And performed a detailed theoretical derivation of the hand-eye-target calibration of the detection system.using the external pa-rameters of the camera to detect the hand-eye-target calibration of the system.Finally,the MATLAB algorithm based on Li group theory is used to solve the opponent-eye-target cal-ibration relationship.Third,adopt 4 image processing methods to improve the sharpness of corner infor-mation of checkerboard target images,the image is subjected to illumination balance,Gaussian filter denoising,maximum inter-class difference method binarization and small area filtering denoising processing,which effectively shortens the image calculation time,improves the clarity of the image of the checkerboard corner point,and improves the clarity and improve the efficiency and accuracy of the camera calibration and hand-eye-target cal-ibration.Fourth,make a detail study of the calibration scheme of robot geometric parameters based on spatial position constraint method and joint axis detection method based on visual inspection.The geometric parameter calibration of the robot is completed by the actual po-sition detection of the axis and the numerical optimization of the geometric parameters,thereby improving the accuracy of the end position of the robot.The spatial position con-straint method detects the position of the robot end in space using visual inspection,com-pares the detected data with the position data in the robot controller to determine the position error of the robot end in space,and then uses the error to establish the error in combination with the kinematic equation of the robot.Then the genetic algorithm is used to optimize the geometric parameters of the robot,and obtained the optimal geometric parameter values under a set of current constraints,finally,the geometric parameter calibration task is com-pleted.The axis detection method detects the position data of the optical center of the cam-era attached to the end of the robot in the rotation process by rotating a single joint axis and adopting a camera calibration technique combined with a checkerboard target,and then us-ing the detection data to perform spatial circle fit,determine the spatial position data of the single joint axis,and then calibrate the geometric parameters according to the joint axis position.Fifth,The experiment verifies the feasibility of the two calibration methods and the validity of the calibration results.accurate calibration of the hand-eye-target position of the detection system is achieved by standard robot pre-calibration,and solved the specific rela-tionship.Then centrally processed the experimental data,and the detailed data results of the spatial position constraint method and the axis detection method are calculated respectively.the specific numerical values of the geometric parameters of the two schemes are obtained.The theoretical research and experiment results shows that the axis detection method used in the study successfully realizes the detection of the position of the joint 1,2,and 3axes,calibrates a₁,a₂,d₁ three parameters and improve the end position accuracy of the robot from 2.350mm to 1.477mm;The space position constraint method successfully realizes the calibration of the optimal value of the geometric parameters of the robot,and improves the end position accuracy of the robot from 2.350mm to 1.524mm.The experimental results fully prove the effectiveness and feasibility of the two calibration methods based on visual detection parameters.