AUTHOR=Feng Lei , Zhang Lianmei , Gao Zepu , Zhou Ruoyun , Li Lan 

TITLE=Gabor-YOLONet: A lightweight and efficient detection network for low-voltage power lines from unmanned aerial vehicle images

JOURNAL=Frontiers in Energy Research

VOLUME=Volume 10 - 2022

YEAR=2023

URL=https://www.frontiersin.org/journals/energy-research/articles/10.3389/fenrg.2022.960842

DOI=10.3389/fenrg.2022.960842

ISSN=2296-598X

ABSTRACT=Maintaining and monitoring low-voltage overhead power lines are of great importance as they are the key link between power system and low-voltage power users. At present, few detection network can be detected accurately on intelligent edge identification devices due to the complex backgrounds and limited characteristics in unmanned aerial vehicle Images as well as the low computing abilities of hardware. In order to give consideration to accuracy and speed, we propose a novel power line detection method, denoted by Gabor-YOLONet, used for intelligent edge identification devices available to UAV.Unlike existing works, the proposed method use Gabor algorithm to extract foreground of power lines from cluttered backgrounds automatically and predict power lines and their auxiliary targets such as insulators in the foreground scene. Besides, we introduce a new inference method which can summarize the average location and orientation of auxiliary targets by clustering to verify the rationality of the predict results for power lines. The experiment results show that our method has the higher accuracy but consumes less computing resources compared with other methods, the mAP of identification for power lines can be 86.6%, and the running time is only 24 ms, with excellent performance on intelligent edge devices.