摘要
线路点云数据结构化是深化专业计算与分析的技术前提。点云分割是点云数据结构化的基础。钢轨作为轮轨走行面,其空间位置的连续及平顺情况直接影响行车安全,因此轨道结构分割中,首先需对钢轨进行分割处理。针对传统欧式距离聚类中线路全景点云数据遍历导致距离阈值难统一、难界定,造成分类过多不易查找,或人工选取初始点及调参带来的自动化程度不高的情况,提出基于反射强度的改进欧式距离聚类钢轨点云分割方法。在对轨道结构特性分析的基础上,采用布料滤波算法进行地面滤波,区分地面点与地物点,精简线路点云为轨道结构点云;融合点云反射强度属性,提出提取率概念,确定钢轨高反射强度区间,进行钢轨顶面点云预分割,进而以轨顶面预分割点作为初始点,根据钢轨断面轨头高和轨头宽构造对角线长度来计算距离阈值,由Kd-Tree找到小于轨头距离阈值的点进行欧式距离聚类,实现对轨头凸集点云的分割。多路段钢轨点云分割试验,精确率及召回率均大于90%,说明该方法可行有效。
Data structuring of railway point cloud is a technical precondition for deepening professional computing and analysis.Point cloud segmentation is the basis of point cloud data structuring.The continuity and smoothness of the space position of the rail,as a wheel-rail running surface,directly affect the safety of train operation.Therefore,in the segmentation of track structure,it is necessary to segment the rail first.In response to the difficulty in unifying and defining distance thresholds due to the traversal of cloud data of all scenic spots of routes in traditional Euclidean distance clustering,resulting in too many classifications and difficulty to find,or the low degree of automation caused by manually selecting initial points and adjusting parameters,an improved Euclidean distance clustering rail point cloud segmentation method was proposed based on reflection intensity.Based on the analysis of the characteristics of the track structure,the cloth simulation filtering algorithm was used to distinguish ground shape points from ground object points,and simplify railway line point clouds into the track structure point clouds.The concept of extraction rate was proposed by integrating the reflection intensity attribute of point clouds to determine the value range of rail high reflection intensity to carry out pre-segmentation of rail top surface point clouds.Furthermore,with the pre-segmentation point on the rail top surface as the initial point,the distance threshold was calculated according to the diagonal length of rail head height and rail head width of rail section.The Euclidean distance clustering was carried out by Kd-Tree to find the points less than the distance threshold of the rail head,so as to realize the segmentation of the convex collection point cloud of the rail head.The multi-section rail point cloud segmentation experiments show that the precision and recall rates are greater than 90%,proving the feasibility and effectiveness of the method.
作者
段晓峰
高伟伟
韩峰
DUAN Xiaofeng;GAO Weiwei;HAN Feng(School of Civil Engineering,Lanzhou Jiaotong University,Lanzhou 730070,China)
出处
《铁道学报》
EI
CAS
CSCD
北大核心
2024年第2期114-121,共8页
Journal of the China Railway Society
基金
国家自然科学基金(51568037)。
关键词
钢轨
点云分割
反射强度
欧式距离聚类
rail
point cloud segmentation
reflection intensity
Euclidean distance clustering
