Insulator Defect Detection in Foggy Condition Based on Improved YOLOX Network

TANG Lu; WANG Shuqing; WANG Niantao; YAO Ruotian; WANG Yibo

doi:10.13296/j.1001-1609.hva.2024.03.028

您当前的位置：

首页 >

文章列表页 >

Insulator Defect Detection in Foggy Condition Based on Improved YOLOX Network

Technical Discussion | 更新时间：2025-04-26

- Insulator Defect Detection in Foggy Condition Based on Improved YOLOX Network
- High Voltage Apparatus Vol. 60, Issue 3, Pages: 223-228(2024)
- 作者机构：
  
  1.湖北工业大学电气与电子工程学院，武汉　 430068
  2.武汉大学电气与自动化学院，武汉　 430072
- 作者简介：
- 基金信息：
- DOI：10.13296/j.1001-1609.hva.2024.03.028
  CLC：
- Received：17 October 2023，
  
  Revised：18 December 2023，
  
  Published：16 March 2024
- 稿件说明：
移动端阅览
TANG Lu, WANG Shuqing, WANG Niantao, et al. Insulator Defect Detection in Foggy Condition Based on Improved YOLOX Network[J]. High voltage apparatus, 2024, 60(3): 223-228.
DOI：

TANG Lu, WANG Shuqing, WANG Niantao, et al. Insulator Defect Detection in Foggy Condition Based on Improved YOLOX Network[J]. High voltage apparatus, 2024, 60(3): 223-228. DOI： 10.13296/j.1001-1609.hva.2024.03.028.

摘要

绝缘子作为输电线路中重要的元器件，对绝缘子缺陷的识别和定位在电网巡检方面有重要意义。为了对雾天情况下绝缘子状态进行准确定位识别，改进了YOLOX目标检测网络。首先，为了加快网络的训练速度，改进了激活函数；其次，为了解决小目标样本相较于整体样本较少和样本不均衡的问题，改进了图像增强方法和损失函数；最后，将改进的YOLOX网络在数据集上与4种目标检测网络进行比较。实验结果表明，在雾天数据集上，改进后YOLOX的准确度、平均准确率和每秒处理图片张数都有一定提升，符合无人机巡检的要求。

Abstract

As an important component of transmission line

the identification and location of insulator defects is of great significance in patrol inspection of power grid. For accurately locating and identifying the insulator state in foggy condition

the YOLOX target detection network is improved. Firstly

in order to speed up the network training

the activation function is improved; Secondly

for solving the problem that the small target samples are less than the whole samples and the samples are unbalanced

the image enhancement method and loss function are improved; Finally

the improved YOLOX network is compared with four target detection networks on the data set. The experimental results show that on the foggy data set

the accuracy

average precision and the number of frames per second of the improved YOLOX are improved to a certain extent

which meets the requirements of UAV patrol inspection.

关键词

Keywords

references

Views

202

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

No data

Related Author

No data

Related Institution

No data

⁰