电瓷避雷器Issue(6)：117-126,10.DOI:10.16188/j.isa.1003-8337.2025.06.014
基于改进YOLOv5s的输电线路绝缘子缺陷检测方法研究
Defect Detection Method for Transmission Line Insulator Based on Improved YOLOv5s
摘要
Abstract
Aiming at the low accuracy in identifying insulator defects in transmission lines,the author pro-poses an insulator defect detection method based on improved YOLOv5s.Firstly,the ECA is introduced into YOLOv5s to reduce irrelevant feature interference and enhance the model's ability to extract target features.Next,the K-means++algorithm is employed to perform clustering analysis on the insulator defect dataset,generating anchor box sizes more suitable for insulator defect detection tasks.Finally,the original detection head is combined with ASFF module to adaptively adjust the fusion weights of features across a spectrum of scales,fully utilizing multi-scale feature information to improve the detection accu-racy of the model.The experimental results indicate that compared to the baseline YOLOv5s,the pro-posed method has achieved a 2.2％increase in mAP@0.5 and a 1.9％increase in mAP@0.5∶0.95,which can more accurately locate and identify the targets,verifying the effectiveness of the improved method.关键词
绝缘子缺陷检测/YOLOv5s/注意力机制/K-means++算法/特征融合Key words
insulator defect detection/YOLOv5s/attention mechanism/K-means++algorithm/fea-ture fusion引用本文复制引用
王素珍,葛润东,吕基岳,邓成禹..基于改进YOLOv5s的输电线路绝缘子缺陷检测方法研究[J].电瓷避雷器,2025,(6):117-126,10.基金项目
山东省自然科学基金项目(编号:ZR2020QF101,ZR2021MF024).Project supported by Natural Science Foundation of Shandong Province(No.ZR2020QF101,ZR2021MF024). （编号:ZR2020QF101,ZR2021MF024）
