Research On Weld Tracking System Of Complex Component Based On Visual Servo

The industrial intelligent welding technology generally adopts the robot teaching and reproduction method to realize automatic welding.However,in the welding process of complex components such as automobiles and ships,due to uncertain factors such as variable working positions and installation errors of welding components,the actual welding process is difficult.There is a deviation between the seam trajectory and the teaching trajectory,and the welding torch cannot adapt to the change of the welding seam to adjust the position and attitude to correct the trajectory.Therefore,aiming at the robot autonomous welding of complex components,this thesis studies the welding seam tracking system of complex components with the goal of automatic welding of automobile tailgates.Firstly,according to the composition principle of laser vision welding seam tracking system,the laser vision sensor is selected and the fixture is designed according to the actual work requirements.The conversion formula from the pixel coordinate system of the image to the robot base coordinate system is obtained,and the calibration error of the hand eye is compensated through experiments.Then,in order to solve the problem of strong noise interference in the process of welding seam tracking,a tracking algorithm based on kernel correlation filtering is proposed in this thesis.By using LBP and HOG fusion features,adding target template adaptive update mechanism and establishing a Kalman filter prediction model for the location of weld feature points,the traditional image processing method is used to extract weld feature points and establish In the target tracking area,after the welding starts,the welding seam fusion feature is calculated and imported into the tracking algorithm model to realize the tracking of the welding seam feature points in the subsequent strong noise environment.In order to solve the problem that the welding torch posture cannot be adjusted automatically when welding the welds of complex components,a method of welding seam pose recognition and welding torch adjustment angle calculation method is proposed.At the same time,the position-based visual servo control mode in the welding seam tracking system is studied.By analyzing the kinematic model of the experimental robot,on this basis,the control scheme of the welding seam tracking system is simulated by using Matlab in the Simulink environment.The simulation results show that the control scheme meets the design requirements of the welding seam tracking system and can meet the expectations Weld trajectories are accurately tracked.Finally,in order to test the feasibility of the scheme proposed in this thesis,a welding seam tracking experimental platform was designed and built,a modular welding seam tracking system software was developed,and an offline tracking algorithm comparison experiment was carried out.Curved lap welding seam welding experiment,the experimental results show that the improved tracking algorithm proposed in this thesis can effectively achieve tracking.During the welding process,the average absolute tracking error of the welding seam tracking system is 0.36 mm and 0.46 mm,and the average error of the robot end welding gun is 3°.It can meet the requirements of robot autonomous welding of automobile tail panels in actual situations.