Research On Welding Location And Trajectory Planning Of Robot Based On Machine Vision

At present,welding with robots instead of by hands has become a trend in the field of welding.In some application scenarios of continuous operation,the position of the weldment will change every time,and the welding position on the weldment is not fixed.the method of teaching and playback can not adjust the posture flexibly according to the changes of the environment.An important way to improve the flexibility is to install a visual location system for the robot.In this thesis,based on the existing research results,a robot label welding system based on machine vision is developed for the application scenarios of the label welding of bundled metal bars.It can search for the position of non fixed weldments,select the welding position and 3D reconstruction,and use the visual guidance robot to adjust the welding trajectory flexibly.The main work of this thesis includes:(1)The hardware platform of the label welding system is constructed,which mainly includes the upper computer,industry camera,laser sensor,6-DOF industrial robot,PLC control cabinet,etc.(2)Aiming at the problem of welding location,a multi-sensor information fusion location method based on monocular vision and laser sensor is proposed,and the threedimensional coordinate recovery formula of space point is deduced.The camera was calibrated by Zhang Zhengyou calibration method,and the internal parameter matrix and distortion parameters of the camera were obtained.The hand eye calibration of the robot is carried out by using the chessboard,and the homogeneous transformation matrix from the camera coordinate system to the welding torch coordinate system is obtained.(3)For the adhesion of bars image,a complete corrosion segmentation algorithm is proposed.In this method,the image of the core position of the weldment is retained by the difference of the image before and after corrosion,and the adhesion image is separated.(4)Aiming at the problem of solder joint selection,an optimal solder joint selection algorithm based on Euclidean distance is proposed to realize the selection of solder joint position on the bar section image and lay the foundation for the subsequent 3D reconstruction.(5)For the trajectory of the welding robot,the trajectory planning is performed on the robot,and the vision system will guide the robot to the welding position to complete the label welding.In addition,for the judgment of the contact state between the welding stud and the bar section,a resistance detecting circuit is designed to effectively solve the problem of the virtual welding caused by the positioning error.Finally,the verification experiment is carried out.The experimental results show that the absolute error in the x and y directions is within 0.9 mm and the absolute error in the z direction is within 1.0 mm.The mean square error is σx=0.378,σy=0.223,σz=0.373,the system has high precision and stability,meeting the welding requirements.The whole system has obvious advantages for the recognition of bar section image,the location of the welding position and the trajectory planning of robot,and has a very good application prospect,which lays a foundation for further realizing the automatic welding of the welding robot.