基于城区机载雷达数据的双重滤波算法  

Double-Filtering Algorithm Based on Airborne LiDAR Data of Urban

作　　者：赵倩[1] 田茂义[1] 刘景瑞[1] 李中洲 王耀国[1] 

机构地区：[1]山东科技大学测绘科学与工程学院,山东青岛266590 [2]中国石油大学(华东)地球科学与技术学院,山东青岛266555

出　　处：《测绘科学技术学报》2012年第3期200-203,共4页Journal of Geomatics Science and Technology

基　　金：国家科技支撑计划项目(2011BAK07B01)

摘　　要：机载激光雷达扫描技术能快速且高精度地获取地面点的3维坐标,而激光雷达数据处理的首要任务就是点云的滤波,也即是将地面点和非地面点进行分离。传统的滤波方法大都是基于一定的地形条件或是小规模数据量进行的。针对城区的3维点云处理提出了一种双重滤波方法:先构建三角网,根据三角面片的角度信息过滤出一部分点云,将剩余点划分成规则格网;然后通过移动最小二乘曲面拟合法,将高差大于一定阈值的点滤除,从而获得地面点云。Airborne laser radar scanning techniques can obtain three-dimensional coordinates of ground point quickly in a high accuracy, but the first task of LiDAR data processing is point cloud filtering, that is the sepa- ration of ground point and non-ground point. Traditional filtering methods are mostly based on a certain terrain conditions or a small amount of data. Then, a new double-filtering method for the complex three dimensional point cloud was proposed： building the triangulation, and filtering out part of the point cloud according to the angle of the triangular facets and dividing the rest point into the regular grid; and then filtering the height difference that is greater than a certain threshold with the moving least squares surface fitting, and finally gai- ning ground point cloud.

关 键 词：LIDAR点云 滤波 三角网 最小二乘 移动曲面拟合 

分 类 号：P237[天文地球—摄影测量与遥感]