中文摘要：

LiDAR已成为获取数字地面模型（DTM）的重要手段。结合城市地形较平坦的特点,提出了一种针对LiDAR离散点云数据提取DTM的分步滤波方法,该方法主要通过高程频率直方图获取滤波主要参数,运用趋势面拟合的方法剔除剩余地物点,然后通过克里金法对地面点进行插值,最终得到DTM。将这种滤波方法应用在上海陆家嘴某区域,并与单纯趋势面拟合法的结果进行比较,实验数据表明,该滤波方法能有效的获取城市DTM。

英文摘要：

LiDAR（Light Detection and Ranging） is thought as an important means to obtain Digital Terrain Model（DTM）.As the terrain is flat in metropolis,so a stepped filtering method aiming at DTM sampling from LiDAR point cloud data is given in this paper.The elevation-frequency-histogram is used for capturing the main filtering parameters,the trend surface model is utilized to eliminate non-ground points,and the DTM is generated eventually by Kriging interpolation.Applied to Lu Jia Zui district,Shanghai,this method works more efficiently and effectively,in compare to the trend surface model.