3D Point Cloud Segmentation Algorithm Based on Improved Region Growing

Abstract

Point cloud segmentation is the key link of point cloud data processing, which can provide important information for subsequent surface reconstruction and feature extraction. In the point cloud segmentation algorithm based on regional growth, the selection strategy of seed points and the judgment basis of growth or not are the two factors that have the greatest influence on the segmentation effect. In this paper, the point with the smallest curvature is used as the seed point. According to the density of the point cloud and the number of points fitting the surface adaptively, the normal vector of each point on the surface of the point cloud is calculated. Then, by calculating the mean and standard deviation of the angle difference between the normal vectors of the points in the neighborhood of various sub-points, the confidence interval is set as the condition of regional growth. Experimental results show that the proposed algorithm improves the effect of point cloud segmentation obviously compared with the traditional region growing algorithm.