作　　者： ; ; ; ;

机构地区： 中国科学院遥感应用研究所

出　　处： 《中国体视学与图像分析》 2009年第2期209-215,共7页

摘　　要： 通过激光雷达数据反演林木参数,特别是树高,将极大地推进激光雷达在林业上的应用。本文首先对LiDAR获取的高密度点云进行预处理,生成适合森林参数提取的规则网格数字表面模型,然后利用形态学滤波的方法逐步去掉非地形要素,形成数字高程模型,最后利用数字表面模型减去得到的数字高程模型,可得到正则化数字表面模型,并求出地物的相对高度信息,在林木上就是最终得到所需要的树木平均高度信息。结合真实LH System ALS40数据进行实验,验证了本文方法的可行性。 Forest parameters, especially forest height, can be inverted from Light Detection And was pro- cessed task, Ranging （LiDAR） dataset, which should promote the applications of LiDAR in forestry. To do the hard task, the cloud point data of LiDAR was processed at first, followed by generating regular grid digital surface model （DSM） to extract forestry information, then morphological filtering was used to re- move non-topographic features, finally relative height information of ground features, including the forest height information, were achieved by subtracting digital elevation model （DEM） from DSM, called nor- malized DSM （nDSM）. Validation of the experiments was carried out on a dataset collected by LH System ALS40 and the feasibility of this technique was proved.

关 键 词： 激光雷达 树高 形态学滤波 正则化数字表面模型

领　　域： [农业科学] [农业科学] [自动化与计算机技术] [自动化与计算机技术]