基于YOLOv5s的继电保护装置型号自动识别方法研究
Research on Type Automatic Identification Method of Relay Protection Device Based on YOLOv5s
彭祥 1管其杰1
作者信息
- 1. 华东送变电工程有限公司,上海 201803
- 折叠
摘要
[目的]在工业自动化和电力系统中,继电保护装置是确保电力系统安全稳定运行的关键组件.为保证设备的正常测试和维护,需要准确识别继电器型号.因此,提出一种基于YOLOv5s的继电保护装置型号自动识别方法.[方法]通过自制数据集,并利用YOLOv5s的高效目标检测能力,结合多尺度特征融合图像处理技术,实现对继电保护装置型号的自动识别.[结果]对系统进行实验测试,结果显示,该方法在继电器型号识别任务中的准确率和召回率分别为90.3%和85.6%,能满足实际应用需求.[结论]该方法不仅提高了设备测试和维护的效率,还为电力系统的安全稳定运行提供了可靠的技术支持.
Abstract
[Purposes]In industrial automation and power systems,relay protection devices are key com-ponents for ensuring the safe and stable operation of power systems.Accurate identification of relay mod-els is crucial for the proper testing and maintenance of equipment.To address this need,this paper pro-poses a relay protection device model identification method based on YOLOv5s.[Methods]By using the highly efficient target detection capability of YOLOv5s and image processing technology such as multi-scale fusion,the automatic identification of the relay protection device model was realized in the self-made data set.[Findings]The experimental results show that the accuracy and recall rates of the method are 90.3%and 85.6%in the task of relay type identification,which can effectively meet the needs of practical applications.[Conclusions]This approach not only enhances the efficiency of equipment test-ing and maintenance but also provides reliable technical support for the safe and stable operation of power systems.
关键词
继保护装置/YOLOv5s/目标检测/型号识别/深度学习Key words
relay protection device/YOLOv5s/object detection/model identification/deep learning引用本文复制引用
出版年
2024
NETL
NSTL
万方数据