Abstract:

Focusing on the problems of occlusion and holes in view planning research,a new method was presented, which could eliminate effectively the problems of occlusion and holes with establishing minimum contain box of the hole boundary point clouds. The whole process of view planning was divided into two parts: the laser scanning view planning of side surfaces and the scanning view planning of hole and occlusion surfaces. The parallel surface of least-square plane about hole boundary points was used as the movement plane of vision system scanning hole surfaces,so the translational motion range of the vision system was greater than the plane range of the minimum contain box,and the point cloud data of hole surfaces were obtained. Finally, experimental results show the effectiveness of this method.  

Key words: machine vision, view planning, minimum contain box, occlusion and hole

摘要：

针对视点规划研究中的遮挡与孔洞问题,提出了通过建立点云孔洞边界点最小包容盒来有效消除遮挡孔洞的新方法。把视点规划分为侧面信息激光扫描视点规划和遮挡孔洞表面扫描视点规划,以孔洞边界点最小二乘平面的平行面为视觉系统扫描遮挡孔洞表面的运动平面,视觉系统的平移运动范围大于最小包容盒的四周平面范围,对孔洞轮廓表面进行点云数据采集。最后通过实例验证了所提方法的有效性。

关键词: 机器视觉, 视点规划, 最小包容盒, 遮挡孔洞