摘要
为实现既有线路中心线的拟合(增改线的设计施工)以及钢轨点云建模(分析钢轨的几何形位),提出一种基于线路特性的铁路钢轨点云提取分割的方法。首先利用点云数据中包含的强度信息以及高速铁路纵坡的限制,通过设置阈值对线路其他结构物进行过滤,再对钢轨点云进行第一次提取(提取后的点云存在大量离群点);然后基于离群点与钢轨点云的空间位置关系,采用统计滤波算法过滤掉绝大部分离群点;最后采用欧式聚类算法将左右股钢轨单独提取出来。为验证该方法的有效性,选取一段路堑式路基和一段隧道作为试验对象,在提取准确度方面,两段试验线路分别为96.85%,96.94%;在完整度方面,两段试验线路分别为94.39%,94.67%。试验结果表明:该方法简单有效,提取的钢轨点云在准确度和完整度上满足要求。
As the most basic component of railway line structures,rail is also the basis of horizontal,high and low,rail direction and other alignment parameters.In order to realize the fitting of existing line center line(design and construction of extension line)and the modeling of rail point cloud(analysis of rail geometry),it is necessary to extract rail point cloud from the railway line.Therefore,a method of rail point cloud extraction and segmentation based on line characteristics was proposed,including the material difference of different line structures and railway line characteristics.Firstly,the intensity information contained in the point cloud data and the restriction of the longitudinal slope of high-speed railway were used to filter other structures of the line by setting thresholds,and the rail point cloud was extracted for the first time.This step greatly reduced the amount of calculation,but there were a lot of outliers in the extracted point cloud.Then,most outliers were filtered by statistical filtering algorithm based on the spatial position relationship between outliers and rail point cloud.Finally,the left and right rails were extracted separately by using euclidean clustering algorithm.In order to verify the effectiveness of this method,a section of line point cloud of cut roadbed and a section of line point cloud of tunnel were selected as test objects.In terms of accuracy,the two sections of test line were 96.85%and 96.94%respectively,and in terms of integrity,the two sections of test line were 94.39%and 94.67%respectively.The experimental data shows that,the method is simple and effective,and the extracted rail point cloud meets the requirements of accuracy and integrity.
作者
杨陆冰
段晓峰
Yang Lubing;Duan Xiaofeng(School of Architecture and Construction,Lanzhou Jiaotong University,Lanzhou 730070,China)
出处
《铁道勘察》
2022年第5期42-46,共5页
Railway Investigation and Surveying
关键词
高速铁路
线路特性
钢轨
线路点云
滤波
分割
high-speed railway
alignment features
rail
line point cloud
filtering
segmentation
作者简介
杨陆冰(1994—),男,硕士研究生,E-mail:236750514@qq.com;通信作者:段晓峰(1978—),女,2000毕业于兰州交通大学道路与铁道工程专业,工学博士,副教授,E-mail:774196620@qq.com。
