With a view to satisfying the robot's requirements for weld grinding,an improved algorithm for the weld seam skeleton extraction based on point cloud data was proposed.Firstly,having the octree-based voxel filtering used to downsample original point cloud of the welding seam;and then,having both random sampling and WLOP operator adopted to obtain initial skeleton points,including having the weld seam skeleton extracted by iterative contraction and having the repulsive function introduced to avoid the genera-tion of wrong branches;finally,having the skeleton lines connected by fitting curves.The experimental re-sults show that,the algorithm can adapt to many different welds,extract the weld seam skeleton with good quality and correctly express the main features of welds.
关键词
点云处理/骨架提取/焊缝特征提取/焊缝磨削
Key words
point cloud processing/skeleton extraction/welding seam feature extraction/weld seam grinding
维普
万方数据