自适应法向量约束的工件分割方法

针对目前对工件点云分割中出现的过分割、欠分割等问题,提出一种结合欧式聚类和自适应法向量约束的工件分割方法.首先对采集到的点云进行体素滤波,使用随机采样一致性(random sample consensus,RNASAC)算法滤除背景并通过统计滤波器去除噪点,接着利用欧式聚类算法将目标点云划分成点云簇,最后针对点云簇中含...

Adaptive normal vector constrained method for workpiece segmentation

Aiming at the problems of over-segmentation and under-segmentation in the point cloud segmentation of the workpiece, a method combining Euclidean clustering and adaptive normal vector constraint was proposed. Firstly, the point cloud of collected was filtered by voxel, the background was filtered by random sample consensus (RNASAC) algorithm, and the noise was filtered by statistical filter, secondly, the target point cloud was divided into point cloud clusters by European clustering algorithm, finally, focused on point cloud clusters containing multiple artifacts of point cloud, the average normal vector of the point cloud cluster was calculated, and different point cloud clusters were segmented by region growth algorithm with different normal vector constraints. The results show that the segmentation method proposed in this paper can effectively segment the workpiece of interest, and the segmentation effect is relatively stable, which provides an experimental basis for the subsequent size measurement of the workpiece.