Acquisition Of Welding Path And Trajectory Planning Based On Passive Vision Sensing

Arc welding robots based on teaching-playback way are widely applied in the industrial production. These robots get welding path by artificial teaching and online programming, which require stable working conditions. These robots fail to deal with the environmental factors and need to be re-teached when the welding workpieces are replaced. This thesis researches the acquisition of welding path and trajectory planning based on vision sensing, which can overcome the shortcomings of teaching-playback robots. This research is of great significance to improve the intelligence level and application value of robots.We build our experimental platform by utilizing an ABB IRB4400robot and a CMOS industrial camera. After obtaining the image of welding seam and image preprocessing such as setting region of interest, graying and smoothing, we make use of various detection operators to detect the image edge, and choose Canny operator as the ultima operator according to the effects of the detection. Combined Hough Transform with least square method to fitting the straight lines and ellipse, we get the functional equation of welding seam under the image coordinate system. We apply checkerboard plane calibration method to precisely calibrate the camera. According to the internal as well as external parameters of camera obtained by camera calibration and the transformational relation of coordinate systems, we get the welding seam equation set under the base coordinate system of robot. Finally, we use spline interpolation method to complete the trajectory planning of robot. The acquisition of welding path and trajectory planning based on passive vision sensing proposed by this thesis can save the computing time of robots, reduce the deviation of welding, and satisfy the requirement of high quality and accuracy of production.Based on the significance and current status of research on passive vision sensing and acquisition of welding path, this thesis focuses on weld image processing, camera calibration experiment and trajectory planning based on spline interpolation method.