Ground Segmentation Algorithm Based on Point Cloud Data

Abstract

Aiming at the problem of under segmentation between ground point cloud and multi-target object point cloud, a ground segmentation algorithm for point cloud data based on multi-region segmentation is proposed in this paper. In the process of LIDAR point cloud filtering, the method of mapping 3D point cloud to grid map is used to detect and filter the ground point cloud. The algorithm uses plane fitting to detect the ground point cloud in each grid, and filters the ground point cloud in the grid to retain the non-terrestrial point clouds. The algorithm can reduce the possibility of under segmentation of ground point cloud effectively, reduce the amount of calculation of subsequent target detection, and improve the accuracy.