A Framework of Power Pylon Detection for UAV-based Power Line Inspection

Regular power line inspection is of great importance for safe and reliable electric power transmission. Compared with the traditional manual inspection, the use of unmanned aerial vehicle with a camera has the advantages of fast speed, low labor cost, and small personnel risk. In order to automatically select the images that may contain faulty power pylons, we introduce a power pylon detection framework by fusing multi-source information, including camera calibration, power pylon model projection and clustering, and feature extraction and matching. Furthermore, we propose an implementation of this framework and evaluate its effectiveness on real power pylon images. The results show that our method is able to detect the power pylon automatically and accurately, and report the abnormal status of power pylon such as missing poles, which indicates the validity of the proposed method.