基于车载点云的道路边界精细提取方法  

Fine extraction method of road boundary based on vehicle-mounted point clouds

作　　者：王泽华 WANG Zehua(Zhejiang Academy of Surveying and Mapping,Hangzhou,Zhejiang 311100,China)

机构地区：[1]浙江省测绘科学技术研究院,浙江杭州311100

出　　处：《北京测绘》2024年第12期1699-1704,共6页Beijing Surveying and Mapping

基　　金：国家自然科学基金(42261074)。

摘　　要：本文针对现有道路边界信息提取存在精度低、效率差等问题,提出了一种基于车载点云道路边界提取方法。该方法实现道路边界的步骤为:首先,为提升后续算法处理效率,采用梯度滤波与布料模拟滤波(CSF)算法分离地面点、非地面点,得到道路边界点候选数据集;其次,采用开源街道地图(OSM)辅助道路边界候选点分段,并通过随机一致性(RANSAC)算法实现道路边界点提取;最后,计算边界断点区域累计曲率值与距离,实现路口判断,对于断点,采用二次多项式曲线填补拟合。采用车载点云数据进行试验,结果表明,本文方法提取道路边界的精度均能超过80%,优于对比方法,证明了本文方法的可行性与优越性。本文研究可为城市道路信息获取提供一定技术参考。This paper proposed a road boundary extraction method based on vehicle-mounted point clouds to address the issues of low accuracy and poor efficiency in existing road boundary information extraction.The steps of extracting road boundaries by using this method were as follows:Firstly,to improve the processing efficiency of subsequent algorithms,gradient filtering and cloth simulation filtering(CSF)algorithms were used to separate ground and non-ground points and obtain candidate datasets of road boundary points.Secondly,open street maps(OSMs)were used to assist in segmenting road boundary candidate points,and the road boundary point extraction was achieved through the random sample consensus(RANSAC)algorithm.Finally,the cumulative curvature value and distance of the boundary breakpoint area were calculated to achieve intersection judgment.For breakpoints,a quadratic polynomial curve was used to fill in the fitting.The experiment was conducted by using vehicle-mounted point cloud data.The results show that the accuracy of the proposed method in extracting road boundaries can reach over 80%,which is superior to the comparison method.This verifies the feasibility and superiority of the proposed method,and the research in this paper can provide certain technical references for obtaining urban road information.

关 键 词：车载激光点云 边界提取 随机抽样一致性(RANSAC) 布料模拟滤波(CSF) 

分 类 号：P237[天文地球—摄影测量与遥感]