Study on segmentation algorithm with missing point cloud in power line

Power line point cloud segmentation and reconstruction is one of the important tasks of airborne lidar (LiDAR) power patrol. In this paper, in view of the fact that the insulator point cloud is difficult to segment, the cross-line point cloud is difficult to deal with, and the existing algorithm model is incomplete, a power line segmentation algorithm is proposed in this paper. In the first step, the rough classification of point clouds is carried out through the grid characteristics of transmission lines, and in the second step, the random consistency algorithm is used to refine the point clouds, and the filtering algorithm considering cross lines is used to accurately extract the point clouds of power lines. According to the local linear model of the power line, the insulator is segmented. Finally, the power line key points are extracted by the local weighted centroids, and the power line reconstruction is carried out. Through practical engineering experiments and applications, the accuracy of power line segmentation is 95.3%, which verifies the applicability and robustness of the proposed algorithm.