Research on Defect Identification of Glass Insulators in Transmission Lines Based on Target Detection and Insulator String Trajectory Fitting
This paper presents a method for identifying glass insulator burst in transmission lines based on target detection and insulator string trajectory fitting. The color image and distance information of the insulator string are obtained by the depth camera. The position of the center point of each insulator is obtained by ellipse detection, and the track line of the center point of the insulator is obtained by fitting. In addition, the location of the target frame of the connecting plates on both sides of the insulator string is obtained through target detection, and the boundary of the track line of the center point of the insulator is obtained as a constraint. Each coordinate point is matched with the distance information of its corresponding point to obtain a discrete sequence of distance values. After outlier removal, interpolation and smoothing processing, the number of minimum points in the sequence is taken as the number of insulator pieces, so as to determine whether there is a phenomenon of insulator burst. The above method can accurately and effectively determine whether the glass insulator string has a flake explosion phenomenon, and can be applied to the power visual inspection equipment (including drones, image and video monitoring devices, etc.) and its analysis background to quickly discover and locate the problem, which helps the operation and maintenance personnel of the power company quickly eliminate the anomaly of the glass insulator string, reduce the operation risk of the power grid, and reduce the power outage accident.
Authors
- Peiwen Zheng (ORCID: https://orcid.org/0000-0003-2041-5898)
- Yanlu Liu
- Dan Han (ORCID: https://orcid.org/0000-0002-0736-4718)
- Hongwu Dong
Institutions
- China Southern Power Grid (China) (CN)
Publication Details
- Published
- 2025-01-10
- DOI
- https://doi.org/10.1109/iceaai64185.2025.10957587
- Primary Topic
- Industrial Vision Systems and Defect Detection
- Type
- article
- Field-Weighted Citation Impact
- 0.00