基于深度学习和图像处理算法的接触网张力补偿异常检测方法  

Anomaly Detection Method of Catenary Tension Compensation Based on Deep Learning and Image Processing Algorithm

作　　者：占栋[1] 张宝奇[2] 陈海波 黄成亮 李林 Zhan Dong;Zhang Baoqi;Chen Haibo;Huang Chengliang;Li Lin(Southwest Jiaotong University,Chengdu,Sichuan 611756;Power Supply Office of Zhengzhou Railway Bureau,Zhengzhou,Henan 450052;Chengdu Tangyuan Electric Co.,Ltd.,Chengdu,Sichuan 610046)

机构地区：[1]西南交通大学,四川成都611756 [2]郑州铁路局供电部,河南郑州450052 [3]成都唐源电气股份有限公司,四川成都610046

出　　处：《工程技术研究》2022年第7期5-8,共4页Engineering and Technological Research

摘　　要：张力补偿装置是接触网链形悬挂结构的重要组件之一,补偿装置失效时会影响受电弓的取流质量,严重时还会引起打弓、断线等重大事故。目前,虽然铁路相关部门投入了大量人力对补偿装置进行巡查,但是仍存在效率低、工作量大、主观性强等问题。针对张力补偿装置异常检测现状及面临的问题,文章利用安装在检测车车顶的两个线阵相机拍摄张力补偿装置,基于深度学习和图像处理算法提出了一种张力补偿异常检测方法,并在实际线路中进行测试,以期解决线路数据中不同张力补偿装置异常工作的问题。The tension compensation device is one of the important components of catenary chain suspension structure.When the compensation device fails,it will affect the current collection quality of pantograph.In severe cases,it will also cause major accidents such as bow striking and wire breaking.At present,although the railway departments have invested a lot of manpower to detect the compensation device,there are still problems such as low efficiency,heavy workload and strong subjectivity.Aiming at the current situation and problems of anomaly detection of tension compensation device,this paper uses two linear array cameras installed on the roof of the detection vehicle to shoot the tension compensation device,and proposes a tension compensation anomaly detection method based on deep learning and image processing algorithm.The test is carried out on the actual line,in order to solve the problem of abnormal work of different tension compensation devices in the line data.

关 键 词：张力补偿装置 深度学习 目标检测 图像处理算法 阈值分割 

分 类 号：U226.8[交通运输工程—道路与铁道工程]