一种接触网导线覆冰检测和厚度估计的方法  

A Method for Icing Detection and Thickness Estimation of Overhead Contact Wire

作　　者：马浩 MA Hao(China Railway Construction Electrification Bureau Group Third Engineering Co.,Ltd,Gaobeidian Hebei 074000,China)

机构地区：[1]中铁建电气化局集团第三工程有限公司,河北高碑店074000

出　　处：《铁道建筑技术》2024年第12期102-104,166,共4页Railway Construction Technology

基　　金：中国铁建股份有限公司科技研究开发计划项目(2021-C44)。

摘　　要：铁路接触网是沿铁路架设为电力机车提供电能的接触网线路,当接触网出现结冰时,机械与电气性能会下降,将影响电力机车受流,产生燃弧现象,烧蚀接触线与受电弓滑板,严重时会造成接触网垮塌、受电弓滑板断裂,引发严重的安全事故。本文提出一种基于深度学习的铁路接触网导线覆冰检测和厚度估计方法,采用深度学习中的实例分割技术实现对铁路接触网导线覆冰的检测,结合标注的先验信息计算覆冰厚度,能够应对复杂变化的场景,且准确性、实时性均满足铁路接触网线路覆冰检测的实际需求。The electrified railway catenary is a catenary line set up along the railway to provide electric energy for electric locomotives.When the catenary is frozen,the mechanical and electrical performance will decline,which will affect the electric locomotive′s flow,produce arc burning phenomenon,and ablate the contact line and pantograph slide board.In serious cases,the catenary will collapse and the pantograph slide board will break,leading to serious safety accidents.The purpose of this paper is to provide a method based on deep learning for railway catenary traverse icing detection and thickness estimation.This method uses the instance segmentation technology in deep learning to realize the icing detection of railway catenary traverse,and calculates the icing thickness by combining the marked prior information,which can cope with complex changing scenes.Moreover,the accuracy and real-time performance meet the actual demand of railway overhead contact line icing detection.

关 键 词：接触网 覆冰监测 厚度估计 图像标注 

分 类 号：U225[交通运输工程—道路与铁道工程]