PTTE:Power Tower Tilt Estimation Algorithm based on LiDAR Point Cloud

Abstract

Power towers are the infrastructure for power transmission in the power grid, and to ensure the safety of power transmission, the state of power towers needs to be inspected regularly. To improve inspection efficiency, this paper proposes a method for detecting the tilt of towers based on the power inspection method of the UAV LiDAR point cloud. The method uses SCF-Net network to semantically segment the power targets in the scene after the UAV has collected the power corridor data, and then uses K-means clustering to obtain individual towers. Then the extracted individual pylons are used to propose a pylon tilt estimation algorithm based on cross-stretcher point cloud plane fitting and normal estimation. Experiments are conducted with the actual collected power corridor data, and the experimental results show that the IOU index of the classification effect of the SCF-Net network on power poles reaches 97.75, and the untilted poles and manually placed tilted poles can be accurately judged. Compared with other methods, the proposed algorithm has a high degree of automation, does not require too much human intervention, and has engineering application value.