扇形格网高程特征图像的车载LiDAR点云道路提取方法  

Vehicle borne LiDAR point cloud extraction of vertical feature images in sector grid

作　　者：冯鹤 FENG He(The Basis Surveying and Mapping Institute of Liaoning Province,Natural Resources Affairs Service Center of Liaoning Province,Jinzhou,Liaoning 121003,China)

机构地区：[1]辽宁省自然资源事务服务中心辽宁省基础测绘院,辽宁锦州121003

出　　处：《测绘技术装备》2024年第2期16-22,共7页Geomatics Technology and Equipment

摘　　要：针对传统高程特征图像道路点提取方法中格网分辨率需手动设定和存储资源冗余问题,本文提出了扇形格网高程特征图像的车载LiDAR点云道路提取方法。该方法采用自适应扇形格网投影,根据激光雷达(LiDAR)设备参数,自动建立可缩放的扇形格网,有效优化了传统方形格网分辨率手动设置,解决了存储资源冗余问题。在该方法的具体实施过程中,首先,依据LiDAR设备参数进行自适应扇形格网投影;其次,获取每个格网内点的高程,运用反距离加权法(IDW)建立高程特征图像;最后,基于高程特征图像设定高程阈值,实施道路点提取。为验证本文方法的有效性,从KITTI数据集中选取不同道路类型的6组数据进行实验。结果表明,道路点提取的最高质量为87.1%,平均质量为80.2%。与传统高程特征图像道路点提取方法相比,本文方法的平均质量高出近3个百分点。For the problems of manual setting grid resolution and redundancy of storage resources by traditional method for road point extraction,a method for vehicle-borne LiDAR point cloud extraction of vertical feature images in sector grid is proposed in this paper.The method uses adaptive sector grid projection to automatically establish a scalable sector grid according to the parameters of LiDAR equipment,and effectively optimizes the manual resolution setting of traditional square grid.In the specific implementation of the method,the adaptive sector grid projection is firstly carried out according to the parameters of LiDAR equipment.Secondly,the elevation of each point in the grid is obtained,and the IDW method is used to establish the vertical feature images.Finally,the elevation threshold is set based on the vertical feature image and the road point is extracted.In order to illustrate the effectiveness of the proposed method,six sets of data of different road types are selected from the KITTI dataset for experiments,and the results showed that the best quality of road point extraction is 87.1%,the average quality is 80.2%,and the average quality of the proposed method is about three percentage higher than the traditional method.

关 键 词：实时点云 自适应扇形格网 反距离加权法 道路点 KITTI数据集 

分 类 号：P232[天文地球—摄影测量与遥感]