面向轨道维护的可变轨距轨道模型与钢轨点云的配准方法  

Registration Algorithm of Variable Gauge Track Model and Rail Point Cloud for Track Maintenance

作　　者：陈丞 张同刚 李涛 沈迅 邓川[4] 金国清 CHEN Cheng;ZHANG Tonggang;LI Tao;SHEN Xun;DENG Chuan;JIN Guoqing(Faculty of Geosciences and Environmental Engineering,Southwest Jiaotong University,Chengdu 611756,China;Infrastructure Section of Huai’an High Speed Railway,China Railway Shanghai Group Co.,Ltd.,Shanghai 200071,China;China Railway Eryuan Engineering Group Co.,Ltd.,Chengdu 610031,China;China Railway Tunnel Group Sanchu Co.,Ltd.,Shenzhen 518000,China;Institute of Surveying,Mapping and Geoinformation,China Railway Fifth Survey and Design Institute Group Co.,Ltd.,Beijing 102600,China)

机构地区：[1]西南交通大学地球科学与环境工程学院,四川成都611756 [2]中国铁路上海局集团有限公司淮安高铁基础设施段,上海200071 [3]中国中铁二院工程集团有限责任公司测绘工程设计研究院,四川成都610031 [4]中铁隧道集团三处有限公司,广东深圳518000 [5]中铁第五勘察设计院集团有限公司测绘与地理信息研究所,北京102600

出　　处：《铁道学报》2023年第1期91-97,共7页Journal of the China Railway Society

基　　金：中国国家铁路集团有限公司科技研究开发计划(P2019G029);中国中铁股份有限公司科技研究开发计划(2019重大-08-01);中铁隧道局科技创新计划(隧研合2018-23)。

摘　　要：基于车载Lidar技术相关的硬件和算法的快速发展,具备发展成快速车载轨道测量技术手段的潜力,从而替代传统的地面人工轨道测量手段。针对其中根据钢轨点云获取轨道轨距、钢轨位置等参数的问题,定义了可变轨距轨道模型,并在此实现可变轨距轨道模型与钢轨点云的配准方法。新算法在配准迭代过程根据钢轨点云到轨道工作边的距离来动态调整模型轨距,从而在轨道配准精度和轨距测量精度两项关键指标获得了同步提高。通过模拟数分析存在不同轨距偏差、超高等情况下算法性能,并和单钢轨轨道模型和固定轨距轨道模型的配准结果进行比较。最后通过一段干线铁路的实测点云进行测试,试验结果表明单钢轨轨道模型配准后左右钢轨的平行性得不到保证;在直线段与固定轨距轨道模型配准精度和轨距测量精度基本相当,配准精度为0.16 mm;在曲线段可变轨距轨道模型配准精度和轨距测量精度不受轨距变化的影响,显著优于固定轨距轨道模型的结果,精度高88.7%。The rapid development of the hardware and algorithm related to the vehicle Lidar technology provides an opportunity to develop a rapid mobile track measurement technology,which is expected to replace the traditional ground manual track measurement.Aiming at the key problems of parameters such as obtaining track gauge and rail position according to rail point cloud extracted from MLS point cloud,a variable gauge track model was proposed,and then a registration method of variable gauge track model and rail point cloud was established to solve the problems.The algorithm dynamically adjusted the gauge of the track model according to the distance between the rail point cloud and the gauge line of the track model in the registration iteration process.Both the registration accuracy and gauge measurement accuracy were improved at the same time.The algorithm performance was analyzed by simulation data in the case of different gauge deviations and superelevation,and the registration results were compared with thoseof single rail track model and fixed gauge track model.Finally,an MLS point cloud of a section of trunk railway was selected to test the proposed algorithm.The experimental results show that the single rail track model cannot guarantee the parallelism of the left and right rails.In the straight line section,both the registration accuracy and the gauge measurement accuracy of the variable gauge track model are almost equal to those of the fixed gauge track model with registration accuracy of 0.16 mm.In the curve section,the registration accuracy and gauge measurement accuracy of the variable gauge track model are not affected by the gauge change,which are significantly superior to the results of the fixed gauge track model,with an accuracy of 88.7%higher.

关 键 词：轨道模型 轨距调整因子 配准 精度 平行性 车载激光点云 

分 类 号：P258[天文地球—测绘科学与技术]