无线电工程2024，Vol.54Issue(6)：1421-1430,10.DOI:10.3969/j.issn.1003-3106.2024.06.010
基于YOLOv5算法的无人机巡检图像中绝缘子单目测距方法
Monocular Distance Measurement Method for Insulator in Unmanned Aerial Vehicle Inspection Images Based on YOLOv5 Algorithm
陈彬 1刘华洲 1李勃铖 2贾燕峰 3郭昊 2孙君录4
作者信息
- 1. 三峡大学 电气与新能源学院,湖北宜昌 443002
- 2. 国网湖南省电力有限公司超高压输电公司,湖南长沙 410100
- 3. 国网河南省电力公司三门峡供电公司,河南三门峡 472000
- 4. 国网河南省电力公司平顶山供电公司,河南 平顶山 467002
- 折叠
摘要
Abstract
In order to enhance the security of UAV flights and improve line operation and maintenance capabilities,a UAV monocular ranging method is proposed for implementing an accurate auxiliary distance measurement function based on the YOLOv5 deep learning algorithm.The existing open source insulator dataset is expanded and calibrated,then the YOLOv5s model is trained,validated and tested to build a distance measurement model for insulator strings,and the design code is added to the detection module to realize the identification and ranging of insulators.The experimental results show that the model can accurately identify different types of insulators on transmission lines and perform accurate ranging,with an average ranging error of 4.76％;the model is most effective in identifying and ranging composite insulators,and the maximum ranging error is 6％when changing the pitch angle and relative distance;the greater the change in brightness under different weather conditions,the greater the error in identifying and ranging insulators,when the brightness increases to 100％,the error can reach up to 12.3％.The average time required for distance measurement is 0.298 4 s.The proposed method enables efficient and highly accurate distance measurement and provides support for safe distance measurement for UAV inspection.关键词
安全距离/绝缘子/深度学习/YOLOv5/图像识别/单目测距Key words
safe distance/insulator/deep learning/YOLOv5/image recognition/monocular ranging分类
信息技术与安全科学引用本文复制引用
陈彬,刘华洲,李勃铖,贾燕峰,郭昊,孙君录..基于YOLOv5算法的无人机巡检图像中绝缘子单目测距方法[J].无线电工程,2024,54(6):1421-1430,10.
