气体绝缘开关设备腔体内部缺陷智能识别与分析系统研究  

Research on Intelligent Identification and Analysis System for Internal Defects in Gas Insulated Switchgear Cavity

作　　者：高辉 佃松宜[1] 郭斌 范智霖 李中萍 钟许可 GAO Hui;DIAN Songyi;GUO Bin;FAN Zhilin;LI Zhongping;ZHONG Xuke(College of Electrical Engineering,Sichuan University,Chengdu Sichuan 610065,China)

机构地区：[1]四川大学电气工程学院,四川成都610065

出　　处：《机床与液压》2025年第5期184-189,共6页Machine Tool & Hydraulics

基　　金：四川省自然科学基金面上项目(2023NSFSC0475)。

摘　　要：气体绝缘开关设备(GIS)腔体内部狭窄、光线不良,人工巡检方式效率低下,易产生漏检。为解决上述问题,开发一套GIS腔体内部缺陷智能识别与分析系统。使用团队研发的GIS检修机器人采集图像;针对GIS腔体内部低照度的问题,提出基于Retinex理论的端到端的低照度图像增强算法,同时提出基于MaskConv的缺陷检测算法——REP-YOLOX,实现了对GIS腔体内缺陷的高精度、高效率检测。在真实的GIS环境内部进行实验,结果表明:该系统能够有效实现对GIS腔体内部烧蚀、螺钉、螺帽和橡胶等4类缺陷的智能识别和分析。The internal cavity of gas insulated switchgear(GIS)is narrow and the light is poor.The manual inspection method is inefficient and prone to missed detection.In order to solve the above problems,an intelligent identification and analysis system for internal defects of GIS cavity was developed.The GIS maintenance robot developed by the team was used to collect images.Aiming at the problem of low illumination inside the GIS cavity,an end-to-end low illumination image enhancement algorithm based on Retinex theory was proposed.At the same time,a defect detection algorithm REP-YOLOX based on MaskConv was proposed,by which the high-precision and high-efficiency detection of defects in the GIS cavity was realized.Experiments were carried out in a real GIS environment.The results show that the system can be used to effectively realize the intelligent identification and analysis of four types of defects such as ablation,screw,nut and rubber in the GIS cavity.

关 键 词：气体绝缘开关设备 内部缺陷检测 图像增强 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]