Anti-vibration hammer detection in UAV image

• Published in: 2017 2nd International Conference on Power and Renewable Energy (ICPRE) pp. 204 - 207
• Main Authors: Heng, Yang, Tao, Guo, Ping, Shen, Fengxiang, Chen, Wei, Wang, Xiaowei, Liu
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.09.2017
• Subjects: anti-vibration hammer; Convolution; deep learning; defect detection; Feature extraction; Inspection; Machine learning; R-CNN; Training; unmanned aerial vehicle (UAV); Unmanned aerial vehicles; Vibrations
• DOI: 10.1109/ICPRE.2017.8390528

Anti-vibration hammer is a key component in the power power line to prevent vibration of power line for a long time, resulting in power line damage, broken and other hazards. Therefore, it is important to accurately locate the position of the anti-vibration hammer in an image for the subsequent judgment, repair or replacement work. However, the traditional anti-vibration hammer positioning and identification is time-consuming, costly and inefficient. An accurate and efficient method is proposed to detect anti-vibrations in aerial images. This method use deep learning to learn insulators by the convolution neural network, and identify and locate the anti-vibrations in aerial images. The proposed algorithm is tested on lots of aerial images and experimental results show the proposed algorithm detected anti-vibration successfully and efficiently. It is promising to significantly facilitate the visual inspection and automatic identification of anti-vibration in aerial images.