Power Equipment Segmentation of 3D Point Clouds Based on Geodesic Distance with K-means Clustering

In the process of substation 3D modeling, it is necessary to cluster and segment the point cloud in order to extract effective points for 3D model reconstruction of power equipment. To solve wrong segmentation problem, this manuscript proposes an improved adaptive k-means clustering. Specifically, firstly, based on the position distribution of the point clouds, the points that are farther away are...