机载LiDAR数据的城区树木点提取方法_程晓光.pdfVIP

网络出版时间：2013-05-23 10:14 网络出版地址：/kcms/detail/11.4415.P1014.005.html 机载LiDAR 数据的城区树木点提取方法 程晓光，黄先锋，张帆 （武汉大学测绘遥感信息工程国家重点实验室，武汉 430079） 【摘 要】在城区的机载LiDAR 点云中一般存在大量打在树木上的点,从点云中提取的树木 点可以应用于城区绿化面积和树木参数的估计，以及树木的建模。针对城区环境，本文在 综合分析树木点和其他地物空间分布模式的基础上，提出了一种只利用点云的几何性质， 结合点云空间分布模式来提取树木点的方法。实验表明，该方法可以取得很高的分类精度， 卡帕系数为0.9713。 【关键词】机载LiDAR；点云分类；树木；空间分布模式 【中图分类号】TP75 【文献标识码】A 【文章编号】1009-2307(2014)02- A spatial distribution pattern-based urban tree points extraction method in airborne LiDAR Abstract : A large amount of points are located on trees in urban airborne LiDAR point clouds. Tree points extracted from point clouds could be used for the estimation of green area and tree parameters and also modeling of tree. At present, the research of tree points extracting method is limited. Based on a comprehensive analysis of spatial distribution pattern of tree points and other objects, a tree point extraction scheme for urban environment was proposed in this paper which only employs geometric metrics and incorporates spatial distribution pattern of point clouds. Experiments proved that this method could achieve high classification accuracy with Kappa coefficient 0.9713. Key words ：airborne LiDAR; point classification; tree; spatial distribution pattern CHENG Xiao-guang, HUANG Xian-feng, ZHANG Fan (State Key Laboratory of Information Engineering in Surveying, Mapping and Remote Sensing, Wuhan University, Wuhan 430079, China)