LIDAR点云数据的建筑物特征线提取  

The extraction of buildings' roof feature lines based on point clouds

作　　者：孟峰[1] 李海涛[1] 吴侃[2] 

机构地区：[1]中国测绘科学研究院,北京100039 [2]中国矿业大学环境与测绘学院,江苏徐州221008

出　　处：《测绘科学》2008年第5期97-99,108,共4页Science of Surveying and Mapping

基　　金：地理空间信息工程国家测绘局重点实验室开放研究课题(编号:200601)

摘　　要：传统的房屋建模主要采用航空航天立体影像来获取房屋的3维数据,通过在数字摄影测量系统上进行立体量测来实现。这种方式费时、费力、周期比较长,影响了成果的现势性。机载激光雷达可直接、快速获取大面积地表点的三维坐标信息,为建筑物特征线提取和建模带来了便利。以往基于LIDAR点云数据进行建筑物特征线提取,大多只是针对简单的平顶房屋和人字型房屋。本文提出了一种能够精确提取建筑物特征线的方法,通过建筑物区域确定、建筑物顶面分类,能够实现诸如平面、斜面和多层次建筑物特征线提取的方法。The traditional method of building reconstruction is mostly realized by the digital photogrammetric workstation. LiDAR can acquire large areas high resolution DSM directly and rapidly. It is beneficial to the extraction of buildings＇ roof feature lines and building model reconstruction. In the past, people just can extract the simple flat and slope roofs＇ feature lines based on point clouds generally. This paper proposes an algorithm, which can extract the roofs＇detail feature lines. By ascertaining the building land, classifying the roofs, it can extract the flat, slop and multi -level buildings＇feature lines.

关 键 词：特征线 DSM影像 顶面分类 等高线追踪 平面相交 

分 类 号：TP79[自动化与计算机技术—检测技术与自动化装置]