An image-based method to classify power lines in lidar point-clouds

Abstract

Power line management becomes critical as power companies need to assure the reliability of their services. Many of them rely on LiDAR scanning of their assets to get information about the status of the power line corridor and possible dangers in the area. In this paper, a novel method to classify wire and pylon points from a LiDAR point cloud is introduced. The method generates images from different measurements of the data to select candidate areas and then applies a clustering algorithm to group and filter out false positives. A prior classification of ground points is not necessary for the method to work. Tests have been conducted on a set with 25 point cloud files to show the effectiveness of the presented method.