有监督学习下的市域铁路隧道结构裂缝边缘识别方法  

A Detecting Method for Structure Crack Edges of Municipal Railway Tunnels under Supervised Learning

作　　者：张琨 陈定方[1] 高铭鑫 黄永亮 刘洋 门燕青[4] ZHANG Kun;CHEN Dingfang;GAO Mingxin;HUANG Yongliang;LIU Yang;MEN Yanqing(Institute of Intelligent Manufacturing and Control, Wuhan University of Technology, Wuhan, 430063;China Railway Siyuan Survey and Design Group Co.,Ltd., Wuhan, 430063;School of Transportation Science and Engineering, Harbin Institute of Technology, Harbin, 150090;Jinan Rail Transit Group Co.,Ltd., Jinan, 250014)

机构地区：[1]武汉理工大学智能制造与控制研究所,武汉430063 [2]中铁第四勘察设计院集团有限公司,武汉430063 [3]哈尔滨工业大学交通科学与工程学院,哈尔滨150090 [4]济南轨道交通集团有限公司,济南250014

出　　处：《中国机械工程》2021年第4期446-453,共8页China Mechanical Engineering

基　　金：山东省交通运输厅科技计划(2019B09_2)。

摘　　要：裂缝病害的定期巡检工作是保证市域铁路隧道结构正常运营的关键。针对市域铁路隧道结构表观裂缝的有效识别问题,提出了有监督学习下的市域铁路隧道结构裂缝边缘识别方法。首先,利用消费级数码相机获取了市域铁路隧道结构表观状态的图像数据;然后,基于Canny算子的初步边缘目标检测结果,构造了边缘目标的几何特征矢量,并在此基础上提出了基于几何特征数据的裂缝边缘识别方法;最后,采用实际市域铁路隧道结构的表观图像数据验证了所提出方法的有效性。Regular inspection of crack diseases was the key to ensure the normal operation of metro tunnel structures.To address this issue,a method for crack detection of metro tunnel structures under supervised learning was proposed herein.Firstly,the images of the metro tunnel structures were obtained by a consumer digital camera.Secondly,the geometric feature vectors of the edge targets were constructed based on the preliminary detection results by Canny operator,and then the crack edges of metro tunnel structures were detected by using the geometric feature vectors of edge targets.Finally,the effectiveness of the proposed method was verified by the image data of the metro tunnel structures.

关 键 词：市域铁路隧道结构 裂缝识别 监督学习 图像处理 主成分分析 

分 类 号：U231.92[交通运输工程—道路与铁道工程]