Electric insulator detection of UAV images based on depth learning

Electric insulators as an indispensable device for electric power networks, maintaining its safe operation is of vital importance. Due to the large number of insulators and wide distribution, the insulator state detection based on aerial images has important practical significance. Insulator images are usually acquired by artificial or aerial collection, at a specific angle, focal length and complex background. For the labor-detection, low detection efficiency, higher detection cost and other interference, an efficient and accurate method is proposed to detect kinds of electric insulators in unmanned aerial vehicle (UAV) images. This method is based on deep learning, learning insulators characteristics through the convolution neural network in complex aerial images, and then to identify a variety of insulators. The proposed algorithm is tested on a diverse set of UAV imagery. Experimental results show that the proposed algorithm can detect electric insulators efficiently and perform better than other electric insulators detection methods. The proposed method is promising for the change detection of the electric insulators.