Transmission line fault detection based on YOLOv3

Qiongqiong Li; Guochu Chen

doi:10.1117/12.2620412

29 November 2021 Transmission line fault detection based on YOLOv3

Qiongqiong Li, Guochu Chen

Proceedings Volume 12080, 4th International Symposium on Power Electronics and Control Engineering (ISPECE 2021); 120804D (2021) https://doi.org/10.1117/12.2620412
Event: 4th International Symposium on Power Electronics and Control Engineering (ISPECE 2021), 2021, Nanchang, China

Abstract

In order to ensure the smooth operation of power system, intelligent monitoring of transmission lines is necessary. At present, there are few public training pictures in the electric power industry, and the detection accuracy is not good. In this paper, a transmission line fault detection method based on YOLOv3 is proposed. This method uses YOLOv3 model and makes corresponding improvements to the network: 1. Collect and simulate foreign body suspension faults of transmission lines in multiple scenarios and sort them into training pictures. 2. Use the regression loss function GIoU to improve the positioning accuracy of the model. 3. Redesign anchor size by using k-means ++ clustering algorithm to accelerate the reasoning process and further improve network detection performance. Through experiments, the average accuracy of the proposed algorithm in transmission line foreign body detection can reach 98%, and the average detection time of the improved network is 0.025s, which is better than the YOLOv3 algorithm in both accuracy and time.

Citation Download Citation

Qiongqiong Li and Guochu Chen "Transmission line fault detection based on YOLOv3", Proc. SPIE 12080, 4th International Symposium on Power Electronics and Control Engineering (ISPECE 2021), 120804D (29 November 2021); https://doi.org/10.1117/12.2620412

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available