Autonomous Localization and Navigation for Transmission Line Inspection of UAV

Abstract

Transmission lines bear incomparable responsibilities for power transmission. In order to prevent transmission line failures, regular inspection is required. Unmanned aerial vehicle (UAV), an intelligent, efficient, and safe inspection method, has solved many disadvantages of manual inspection, robot, and helicopter inspection. At present, GPS is utilized for positioning and tracking of UAV during inspection. However, it is necessary to find a more convenient and reliable positioning method in such place where there is no GPS signal. Therefore, UAV is combined with binocular vision to conduct transmission line inspection. For the line tracking during inspection, the line features extracted from a single frame are converted into point features in the parameter space of continuous frames. And a two-layer image pyramid hierarchical matching structure is designed to improve the algorithm efficiency. Considering about the demand for autonomous landing of UAV, the visual recognition and spatial pose acquisition algorithm of the cooperative target capable of visual guidance are studied. An experiment for visual tracking of transmission lines in continuous frames is designed. The adjustment of camera attitude is analyzed to verify the accuracy of the designed tracking algorithm. The robustness of the algorithm is verified by the strong and weak light environment respectively. Finally, the autonomous landing experiment of the moving ArUco target is completed. Based on the transmission line positioning and tracking algorithm, the autonomous navigation of UAV is developed, providing a theoretical basis and verification for transmission line inspection under low GPS signal environment.