基于抗差最小二乘配置的点云内插DEM算法研究  

Research on DEM Interpolation Algorithm by Point Clouds Based on Robust Least Squares Collocation

作　　者：李成仁[1] 

机构地区：[1]上海市测绘院,上海200063

出　　处：《测绘与空间地理信息》2015年第12期209-212,215,共5页Geomatics & Spatial Information Technology

摘　　要：针对传统的点云内插DEM算法存在模型误差和点云粗差等问题,提出利用抗差最小二乘配置方法进行DEM内插。该方法可将内插模型误差当作最小二乘配置中的随机信号来处理,并采用基于抗差M估计的协方差函数拟合方法,对粗差数据进行剔除或降权处理,削弱其对协方差函数拟合精度的影响,从而提高DEM内插精度。最后采用三种实验方案对比分析,实验结果表明:抗差最小二乘配置插值方法具有更高的插值精度。Due to the existence of error for the interpolation model and gross error in traditional point clouds interpolation algorithm,robust least squares collocation is introduced to configure the DEM interpolation. This method can handle model error as a kind of random signal in least squares collocation. In this method,robust estimation of covariance functions based on M robust fitting methods is introduced to eliminate gross errors or decrease the weight of them,thus weakening the abnormal elevation values affected accuracy of covariance functions,to improve the accuracy of DEM interpolation. Finally,three kinds of experimental scheme are adapted to comparative analysis,results show that robust least squares collocation interpolation method can get calculation results with higher precision.

关 键 词：机载LiDAR 点云 格网模型 DEM内插 抗差估计 最小二乘配置 

分 类 号：P207.1[天文地球—测绘科学与技术]