结合扫描线与栅格投影的车载LiDAR实时道路边界提取  

Real- time road boundary extraction from vehicle- mounted LiDAR data integrating scanning lines and grid projection

作　　者：陈姝 CHEN Shu(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1期86-92,共7页Geomatics Technology and Equipment

摘　　要：针对车载激光雷达(LiDAR)实时点云数据提取道路边界困难的问题,本文采用扫描线与栅格投影地图相结合的方法进行道路边界提取。首先,基于方位角周期性实施扫描线分离;其次,采用单束点云空间邻域联合分割滤波分离路面点,保留非路面点进行障碍物地图建立;再次,以搜索格网两方向的搜索方式,增加高程极差和路宽约束条件进行道路边界提取;最后,通过多项式曲线拟合算法进行边界拟合,并选取两组代表性KITTI郊区道路数据进行试验。试验结果表明,本文方法能够有效提取道路边界,且应用效果良好,平均准确度可达80%以上。Scanning line and grid projection map are combined in this paper to extract road boundaries from real-time vehicle-mounted LiDAR point cloud data.Firstly,scanning line is separated periodically based on azimuth.Secondly,the single-beam point cloud spatial neighborhood joint segmentation filter is used to separate the pavement points,and the non-pavement points are retained to establish the obstacle map.Thirdly,the search method in both directions of the search grid is used to add the elevation range and road width constraints for road boundary extraction.Finally,boundary fitting is performed by polynomial curve fitting algorithm,and two sets of typical KITTI suburban road data are used to test.The results show that the proposed method can be used to effectively extract road boundaries,it can get good application effect,and the average accuracy can be more than 80%.

关 键 词：扫描线 栅格投影 搜索格网 路坎点云 

分 类 号：P232[天文地球—摄影测量与遥感]