东北电力技术2024，Vol.45Issue(11)：58-62,5.
基于改进YOLOv5的绝缘子污损检测方法
Insulator Fouling Detection Method Based on Improved YOLOv5
张全辉 1赵晋级1
作者信息
- 1. 国网淮南市潘集区供电公司,安徽 淮南 232000
- 折叠
摘要
Abstract
Machine vision based detection of insulator fouling damage is an important method to improve the automation of power sys-tem inspection and inspection efficiency.According to the problems of slow detection speed,high network complexity,and difficulty in achieving accurate detection of insulator fouling damage detection based on deep neural networks,an improved YOLOv5 based insu-lator fouling detection method is proposed.The GhostNet network is used to lightweight design the backbone network of YOLOv5,re-ducing the complexity of the network and improving the detection speed of the model.The CBAM attention mechanism is introduced in-to the feature extraction network to enhance the model's perception ability and improve its detection accuracy.The experimental results show that compared to the YOLOv5 model,the proposed model reduces computational complexity by 71%,parameter quantity by 76%,and detection speed by 31 FPS.At the same time,the detection accuracy can reach 94.6%.Therefore,the proposed model has better performance and can efficiently achieve insulator fouling detection,and provide a certain reference for the maintenance deci-sion-making of power equipment.关键词
绝缘子/污损检测/YOLOv5/CBAM注意力机制Key words
insulator/fouling detection/YOLOv5/CBAM attention mechanism分类
信息技术与安全科学引用本文复制引用
张全辉,赵晋级..基于改进YOLOv5的绝缘子污损检测方法[J].东北电力技术,2024,45(11):58-62,5.
