最小二乘曲面拟合的LiDAR数据滤波方法  

Surface fitting filtering based on least square method from LiDAR data

作　　者：熊俊华[1] 方源敏[1] 邓德标[1] 

机构地区：[1]昆明理工大学国土资源工程学院,昆明650093

出　　处：《测绘科学》2013年第4期74-76,共3页Science of Surveying and Mapping

摘　　要：机载激光扫描技术(LiDAR)可直接获取数字表面模型,但要获得高精度的数字地面模型(DTM)需要滤波去除地物点,这是一项十分复杂和耗时的工作。本文提出一种基于二次多项式的局部最小二乘拟合滤波算法,对数据进行规则格网化处理,对每个格网的数据进行局部最小二乘拟合滤波处理,并缩小格网大小迭代进行滤波。实验结果表明提出的方法能得到很好的效果。Airborne Light Detection and Ranging （LiDAR） technology can construct digital surface model directly, but it is com- plicated and time-consuming to filter surface feature points to make digital terrain model （DTM）. This paper provided a filtering algo- rithm, which is a local least square fitting method based on quadratic polynomial. Firstly, the regular grids were generated. Then these points were filtered in each grid with the algorithm. Finally, the size of grid was reduced and these points were filtered repeatedly. The result of experiment showed that the method could have a good effect.

关 键 词：LIDAR DTM 最小二乘拟合 规则格网 

分 类 号：TP751.1[自动化与计算机技术—检测技术与自动化装置]