机电工程技术2024，Vol.53Issue(8)：197-200,4.DOI:10.3969/j.issn.1009-9492.2024.08.042
基于改进YOLOv5s的绝缘子缺陷检测方法
Insulator Defect Detection Method Based on Improved YOLOv5s
王子玉 1陈佳星 1白博文 1徐爱婷1
作者信息
- 1. 辽宁工程技术大学电气与控制工程学院,辽宁葫芦岛 125105
- 折叠
摘要
Abstract
Insulators are crucial for the safe operation of high-voltage and ultra-high-voltage overhead transmission lines,making accurate identification of insulator images particularly important.To improve the accuracy of insulator defect detection,this study proposes an insulator defect detection method based on an improved YOLOv5s model.By introducing the scSE attention mechanism to achieve calibration in both channel and spatial aspects,the model's ability to identify insulator defects is strengthened,thereby further enhancing the accuracy of insulator image recognition,especially performing more outstandingly in cases of obstruction and similar target interference.Additionally,Mosaic data augmentation processing is added on this basis to provide more data for training the model,enabling the model to perform excellently in multiple scenarios.Experimental results show that the algorithm achieves an accuracy of 92.8%,a recall rate of 97.5%,and a mean average precision(mAP)of 98.8%.Its detection accuracy is significantly superior to that of other models,and it demonstrates outstanding robustness in complex scenarios,providing new insights for insulator maintenance work.关键词
绝缘子检测/YOLOv5s模型/scSE注意力机制/图像处理Key words
insulator detection/YOLOv5s model/scSE attention mechanism/image processing分类
信息技术与安全科学引用本文复制引用
王子玉,陈佳星,白博文,徐爱婷..基于改进YOLOv5s的绝缘子缺陷检测方法[J].机电工程技术,2024,53(8):197-200,4.
