An Outlier Removal Method from UAV Point Cloud Data for Transmission Lines

Abstract

Outlier removal is a crucial preprocessing step in transmission line extraction, inspection, and reconstruction from UAV point cloud data. This paper proposes an automatic and accurate outlier removal method in complex scenarios: First, we select the pylons as the center and specify a threshold value as the radius to roughly extract the pylons; we then apply the European clustering algorithm from the remaining points to obtain the transmission lines and the ground-based objects. Three different denoising algorithms were applied to the three-class of objects. We implement the DBSCAN (Density-based spatial clustering of applications with noise algorithm) to denoise the pylons and the improved DBSCAN algorithm to denoise the transmission lines, and finally the statistical outlier removal method on ground-based objects. The method is validated on the UAV point cloud data and compared with the traditional radius filtering outlier removal algorithm and statistical outlier removal method, whose experimental results show that the proposed method is more robust in complex environments.