Abstract: In view of the problem that the robot grinding trajectory deviates from the actual workpiece weld due to the manufacturing error of the workpiece and the positioning and clamping error in the current robot weld grinding process,resulting in poor consistency of grinding quality,an adaptive correction system of the robot weld grinding trajectory based on the feature matching of the point cloud model was developed. The system uses the surface structure light camera to scan the feature area of the workpiece to obtain the real morphology of the workpiece,matches with the standard point cloud template,and obtains the deviation conversion matrix of the actual workpiece coordinate system,which is applied to the robot master program to realize the rapid guidance of the initial point of welding grinding and the online adaptive correction of the grinding track. Finally,the experimental results show that the system has good performance of automatic correction and stability.

Key words: weld grinding, offline programming, biocular surface structure light, point cloud matching, adaptive correction