中文摘要：

机载激光雷达是一种快速获取高精度三维地理数据的新技术，对其获取的不规则分布的三维点云数据过滤处理是目前的研究前沿，也是机载激光雷达应用的关键技术。本文在分析典型地物激光雷达点云空间分布基本规律的基础上，重点研究了典型地物表面及其边缘的点云空间分布特征及基于TIN结构的邻近点云高程突变规律，并设计了相应的点云过滤算法。论文还对算法的参数选择及相应的误差进行了探讨，并进行了LIDAR数据DEM提取实验。

英文摘要：

LIDAR is a new technology for obtaining directly the highly accurate 3D digital terrain surface, and the filtering of LIDAR point cloud proved a challenge task and still a key technology in the application of airborne LIDAR data. In this article, the distribution characteristic of 3D point cloud from typical objects＇ surface and margin is discussed, and then ,he trait of elevation- jump between points and its neighbors is studied based on the TIN（Triangulated Irregular Network） structure, and then a filtering algorithm for the point cloud extraction through the range of elevation - jump is discussed, and at last, parameters setting and related errors of the algorithm are studied, which followed by a experiment of DEM extraction with the algorithm from LIDAR data. And the result of the experiment shows that the algorithm suit to filter points with obvious elevation - jump.