基于LiDAR点云的建筑物检测方法研究

摘要

提出了一种基于三角面元的LiDAR数据建筑物检测方法.首先对点云数据构建不规则三角网,然后根据三角面元的特征信息对其进行分类,接着利用面元之间的邻接关系对其进行聚类,最后对聚类点云进行跟踪得到建筑物的轮廓.以国际摄影测量与遥感学会(ISPRS)提供的城区LiDAR点云为实验数据进行建筑物检测试验.与以点云或分割块为处理基元的检测方法相比,该方法能够更加准确地提取建筑物轮廓,正确率可达96％,完整率可达85％.%A building detection method based on triangle primitives is proposed.Firstly,TIN is built using the LiDAR point cloud data.Secondly,triangular primitives are classified based on their characteristics.And then,the classified triangular primitives are clustered based on their adjacency.Finally,the clustered point cloud are traced to get the building outline.The proposed method is validated on LiDAR point cloud data of urban area provided by International Society for Photogrammetry and Remote Sensing (ISPRS).Comparing with building detection methods that using point cloud or segmented blocks as primitive,the proposed method is able to extract building outlines more accurately,with correctness up to 96％,and completeness up to 85％.