结合TIN约束与密度聚类的机载LiDAR道路点云提取

针对机载LiDAR道路点云提取过程中自动化提取困难,停车场、水泥地以及与道路相连的地面点难以去除等问题,提出一种三角网约束与密度聚类相结合的机载LiDAR道路点云提取方法。在已有滤波结果的基础上,该方法首先根据道路点云样本的强度信息提取初始道路点,建立Delaunay三角网,运用三角网边长约束精化初始道路点;然后,通过密度聚类算法提取连通性较好且密度较大的独立三角网;最后,采用数学形态学算法优化道路边缘,确定最终道路点。实验选取国际摄影测量与遥感协会提供的两组城市机载LiDAR点云数据进行道路点云提取,结果表明:本文算法可以较好地进行道路点云的自动提取,且对不同类型的道路具有良好的自适应性,验证了算法的可靠性。

An Algorithm Integrating Constrained TIN and DBSCAN for Road Point-clouds Extraction from Air-borne LiDAR

Aiming at the difficulties in automatic road point extraction in places like parking lots,concrete floors,and grounds adjacent to roads,an algorithm integrating TIN and DBSCAN( Density-Based Spatial Clustering of Applications with Noise) is proposed. On the basis of the filtering results,we firstly extract the initial road points according to the echo intensity information of the samples of the roads points,establish Delaunay triangulation network,and refine the initial road points by triangulation length constraint. Secondly,we extract the independent triangulation network which has better connectivity and large density by DBSCAN algorithm. Finally,we optimize the road edges by mathematical morphology algorithm to secure the ultimate road points. In the experiment,we select air-borne LiDAR of two cities from the International Association for Photogrammetry and Romote Sensing to extract roads points. The results indicated that the proposed algorithm can achieve road extraction more accurately and automatically and has better adaptability for different types of roads,which verifies the reliability of the road extraction algorithm.