The Study Of Key Technology Of Visual Servo System Of Welding Robot

With the requirements of welding efficiency and weld quality continue to increasein industrial production, people pay more intention on the work environment and laborintensity. Welding robots is used increasingly in industrial production. Currently theadaptability of welding robot to the work environment is very low so that vision sensoris widely used to solve this problem which has a significance influence on theimprovement of intelligent welding robot.Firstly, in this paper the camera’s the principle and method of calibration isanalyzed, and combines with Calibration Toolbox of MATLAB, and then theestablishment of the image plane coordinate system and code of the extracted referencepoint are optimized, based on the parameters of Bumblebee2to complete the calibrationof binocular camera and the hand-eye system.Secondly, the information of weld image collected by the camera is analyzed, usingtwo one-dimensional Gaussian filter of the vertical direction to smooth the image andthen a gradient histogram image segmentation method is proposed. Extraction of mainedge information is accomplished through this method which is more efficientcomparing to the traditional methods. And then the3D information of the weldaccording to3D reconstruction theory is achievedUltimately, According to the characteristics of the welding robot, the PBVS isapplied, combining the position information through visual sensor with inversekinematics to fulfill the visual servo control. Take the PUMA-560welding robot as anexample, the kinematics model is established by simulink and using some function tooptimize the control parameters. The simulation results show that the accurate seamtracking task is completed.