Transmission power line fault detection based on deep leaning methods

摘要

In order to extract power line and find out faults of power line by overcoming traditional power line fault detection methods, it is proposed to use a deep learning method to detect the power line fault. Deep learning is a powerful method and excels at images recognition and object detection with accuracy. Deep learning methods could be trained end to end. The application of deep learning to detect fault in power line is structure in 3 steps as follows: in the first step uses deep leaning methods combined with sliding approach to make predictions of all part of input image and achieves the output map. In the second step the output map is pre-processed to make it more conducive to localization. Final step is object detection is accomplished according to the information on pre-processed output map.