Multiple Defect Detection of Insulators Based on Improved YOLOv5 Algorithm

WU Zhenliao; WU Zhengping; SUN Shuifa

doi:10.13296/j.1001-1609.hva.2024.12.011

您当前的位置：

首页 >

文章列表页 >

Multiple Defect Detection of Insulators Based on Improved YOLOv5 Algorithm

Research & Analysis | 更新时间：2025-04-26

- Multiple Defect Detection of Insulators Based on Improved YOLOv5 Algorithm
- High Voltage Apparatus Vol. 60, Issue 12, Pages: 95-102(2024)
- 作者机构：
  
  1.三峡大学电气与新能源学院，湖北　宜昌　 443002
  2.三峡大学湖北省输电线路工程技术研究中心，湖北　宜昌　 443002
  3.三峡大学计算机与信息学院，湖北　宜昌　 443002
- 作者简介：
- 基金信息：
  
  National Natural Science Foundation of China(61871258)
- DOI：10.13296/j.1001-1609.hva.2024.12.011
  CLC：
- Received：09 July 2024，
  
  Revised：17 September 2024，
  
  Published：16 December 2024
- 稿件说明：
移动端阅览
WU Zhenliao, WU Zhengping, SUN Shuifa. Multiple Defect Detection of Insulators Based on Improved YOLOv5 Algorithm[J]. High voltage apparatus, 2024, 60(12): 95-102.
DOI：

WU Zhenliao, WU Zhengping, SUN Shuifa. Multiple Defect Detection of Insulators Based on Improved YOLOv5 Algorithm[J]. High voltage apparatus, 2024, 60(12): 95-102. DOI： 10.13296/j.1001-1609.hva.2024.12.011.

摘要

针对绝缘子多缺陷检测精度低、检测速度慢的问题，提出一种改进YOLOv5准确判别绝缘子多缺陷检测算法（YOLOv5-GSEM）。首先通过引入GhostNet结构替换原始网络YOLOv5主干网络C3模块，提升网络运算速度；并在SPPF后引入无参注意力模块SimAM，增强有效特征，抑制干扰特征；其次引入增强特征金字塔网络（EFPN）和多尺度特征融合网络（multiscale feature fusion network，MFFN），充分融合多尺度特征，提升网络对绝缘子多缺陷的检测精度。实验结果表明，文中提出的模型平均精度均值（

mAP

0.5

）达到87.8%，较YOLOv5算法提升了2.7%，检测速度提升了4.6%，该网络的提出为绝缘子多种缺陷检测问题提供一种更有效的方法。

Abstract

Aiming at such issues as low detection accuracy and slow speed of multi-defect detection of insulator

a kind of improved YOLOv5 accurate detection algorithm for multi-defect detection of insulator (YOLOv5-GSEM) is proposed. Firstly

GhostNet structure is introduced to replace C3 module of the original YOLOv5 backbone network to improve the network computing speed. After SPPF

the non-parametric attention module SimAM is introduced to enhance the effective features and suppress the interference features. Secondly

the enhanced feature pyramid network (EFPN) and the multiscale feature fusion network (MFFN) are introduced to fully integrate multiscale features and improve the detection accuracy of multi-def

ects of insulator by the network. The experimental results show that the average of the accuracy of the model (

mAP

0.5

) proposed in this paper reaches 87.8%

which is 2.7% higher than that of YOLOv5 algorithm

and the detection speed is increased by 4.6%. The proposed network provides a more effective method for the detection of various defects of insulator.

关键词

Keywords

references

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

Insulator Defect Detection Algorithm for Multi-scale Feature Extraction Based on Improved YOLOv7

Research on Multi-type Defect Detection of Insulator for Transmission Line Based on GE-YOLO Algorithm

Research and Development of Lightweight 126 kV Gas Insulated Switchgear

Super-resolution Reconstruction of Infrared Image of Transmission Line Insulator Based on Lightweight Residual in Residual Dense Net

Research on Multi Type Defect Detection of Transmission Line Insulators Based on GE-YOLO Algorithm

Related Author

ZHANG Miao

LI Yongling

ZHAO Ruhai

LEI Jingfa

GUO Yihan

SUN Hong

WANG Qiang

FENG Yuntao

Related Institution

School of Mechanical and Electrical Engineering, Anhui Jianzhu University

Key Laboratory of Intelligent Manufacturing of Construction Machinery, Anhui Education Department

Sichuan Provincial Key Laboratory of Process Equipment and Control Engineering

College of Electrical and New Energy, China Three Gorges University

Hubei Provincial Engineering Research Center of Intelligent Energy Technology

⁰