一种改进的LiDAR数据过滤方法  

A progressed filtering method of LiDAR data set

作　　者：黄燕[1] 邓喀中[1] 刘君城[1] 

机构地区：[1]中国矿业大学环境与测绘学院,江苏徐州221008

出　　处：《测绘科学》2010年第5期188-189,91,共3页Science of Surveying and Mapping

摘　　要：Lidar的数据过滤是数据预处理的重要步骤,也是获取高精度数字高程模型的关键。现有很多典型的过滤算法都是基于地面种子点逐渐恢复地面的,比如自适应TIN过滤法、迭代线性预测法,分层恢复过滤法等。初始地面种子点选择的好坏将直接反映到过滤结果的精度上,尤其是对丘陵地区的地形恢复,影响更大。本文对基于种子点的TIN三角网加密法(TS过滤算法)进行了深入剖析,并在种子点选择方法上提出了一种基于移动窗口法和最小残差法的混合式种子点选择方法,有效地提高丘陵地区地形恢复的准确度。Not only the filtering of Lidar data set is one of the important steps in Lidar data preprocess,but also it is the crucial link to obtain the DEM with high precision.There＇s many typical Lidar filtering methods recovering terrain based on ground points seeds,the position and distribution of ground points seeds will directly effect the filtering result,especially in hilly areas.This paper detailedly studied the adaptive TIN filter based on seeds of ground points（TS filter）,brought about some improvement to TS filtering in the aspect of how to choose ground point seeds.The mixed method,which efficiently improved the recovery precision of terrain especially on hills,is the mixture of the window move method and the least residuals method.

关 键 词：雷达 种子点 TS过滤算法 移动窗口法 最小残差法 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]