文档名:基于TLS数据的站场线路点云提取算法
摘要:铁路站场线路几何信息对于铁路安全管理与维护具有重要意义.由于铁路站场内包含多条线路,且轨道错综复杂,使得从大场景点云中自动提取多股道钢轨点云成为难题.地面激光扫描TLS(TerrestrialLaserScanning)作为非接触式测量手段,可快速获取铁路场景中的海量点云数据.针对TLS技术获取的铁路站场点云数据,提出一种基于Delaunay三角网聚类的多股道钢轨点云提取算法.基于分割-归并的思想,在获取铁路站场高精度点云后,沿站场线路方向将点云分为若干段,基于轨道平顺性特征,利用三角网聚类算法逐段提取钢轨顶面点云.在归并阶段整合站场中各股道轨面点云信息,将各段轨面点云连接起来,同时匹配左右轨面点云.将该方法在玉林站部分站场区域进行实例验证,提取到的轨道点云在对象层面上的总体精度为93.95%,完整度为90.57%,准确度为97.59%,相较于平面格网法,提取总体精度提升了5.65%,准确度提升了18.49%.在10处截面提取轨面宽度与轨距,统计结果表明轨面宽度中误差为5.2mm,轨距中误差为5.3mm,满足工程精度需要.实例结果表明,算法可准确有效提取站场多股道钢轨顶面点云,为铁路场景中其他结构物的TLS数据提取工作提供借鉴思路.
Abstract:Thegeometryinformationpertainingtorailroadswithinstationsholdssignificantimportanceforrailroadsafetymanagementandmaintenance.Sincetherearemultiplerailtracksinastation,itisdifficulttoautomaticallyextractpointcloudoftracksfromlargefieldpointcloud.TerrestrialLaserScanning(TLS),asanon-contactmeasurementmeans,canquicklyobtainlargeamountsofpointclouddatainrailroadscenes.BasedonDelaunaytrianglenetworkclustering,apointcloudextractionalgorithmwasproposedtoextractrailtrackpointcloudfromthepointclouddataofarailroadstationobtainedbyTLStechnology.Basedontheideaofsegmentation-merging,themethodsegmentedthehigh-precisionpointclouddataofrailwaystationyardsintoseveralsectionsalongthedirectionoftherailwaytracks,andthenusedthetrianglemeshclusteringalgorithmtoextractthesteelrailtopsurfacepointcloudbasedonthetrackregularityfeatureoftherailwaytracks.Inthemergingstage,therailsurfacepointcloudinformationofeachsectioninthestationyardwasintegrated,andtherailsurfacepointcloudofeachsectionwereconnectedwhilematchingtheleftandrightrailsurfacepointcloud.ThemethodwasverifiedusingdatafromYulinstation,demonstratinganoverallaccuracyof93.95%attheobjectlevel,theintegrityof90.57%,andanaccuracyof97.59%.Comparedwiththeplanargridmethod,theoverallaccuracyofextractionisimprovedby5.65%andtheaccuracyisimprovedby18.49%.Basedontheextractedwidthoftrailsurfacesandgaugesat10crosssections,statisticalresultsindicatethatthestandarddeviationsoftherailsurfacewidthandthegaugeare5.2mmand5.3mmrespectively,meetingtheprecisionrequirementofengineering.Theexampleresultsshowthatthealgorithmcanaccuratelyandeffectivelyextractthepointcloudfromthetopsurfaceofthetracksofmultiplerailroadswithinthestation,whichprovidesareferenceideafortheTLSdataextractionworkinotherstructureswithinrailroadscenes.
作者:方一鹏 宋占峰 李军Author:FANGYipeng SONGZhanfeng LIJun
作者单位:中南大学土木工程学院,湖南长沙410075
刊名:铁道科学与工程学报
Journal:JournalofRailwayScienceandEngineering
年,卷(期):2024, 21(2)
分类号:U216.3
关键词:地面激光扫描 点云 主成分分析 Delaunay三角网 聚类算法
Keywords:terrestriallaserscanning pointcloud principalcomponentanalysis Delaunaytriangulationnetwork clusteringalgorithm
机标分类号:P208U291.1TP391.41
在线出版日期:2024年3月19日
基金项目:国家重点研发计划基于TLS数据的站场线路点云提取算法[
期刊论文] 铁道科学与工程学报--2024, 21(2)方一鹏 宋占峰 李军铁路站场线路几何信息对于铁路安全管理与维护具有重要意义.由于铁路站场内包含多条线路,且轨道错综复杂,使得从大场景点云中自动提取多股道钢轨点云成为难题.地面激光扫描TLS(TerrestrialLaserScanning)作为非接触式...参考文献和引证文献
参考文献
引证文献
本文读者也读过
相似文献
相关博文
基于TLS数据的站场线路点云提取算法 Point cloud extraction algorithm based on TLS data in railway stations
基于TLS数据的站场线路点云提取算法.pdf
- 文件大小:
- 5.23 MB
- 下载次数:
- 60
-
高速下载
|
|