Transmission Line Inspection Image Recognition Technology Based on YOLOv2 Network
Zongqi Mu · 2018
With the gradual improvement of the intelligence of power grid inspection, it is of great significance to apply image recognition technology based on deep learning to the intelligent inspection of transmission lines. This paper proposes an image recognition technology for transmission line equipment based on YOLOv2 network. It utilizes the powerful learning ability and fast recognition speed of YOLOv2 network to accurately learn different characteristics of multiple devices under the condition that the background of transmission lines is varied, and quickly complete the image recognition task. After conducting thorough training on the image of the transmission line obtained by inspection, the accuracy of the YOLOv2 network reached 90.9%, and the recognition speed reached 30 frames per second. The simulation results show that the image recognition technology based on YOLOv2 network balances the recognition accuracy and speed well, and it satisfies the requirements of real-time intelligent inspection.