基于主成分分析法的DEM粗差定位  

Detecting gross errors in DEM of regular data based on PCA

作　　者：杨晓云[1,2] 梁鑫[2] 岑敏仪[1] 顾利亚[1] 

机构地区：[1]西南交通大学土木工程学院地理信息工程中心,成都610031 [2]广西工学院土木建筑工程系,广西柳州545006

出　　处：《测绘科学》2009年第2期128-129,87,共3页Science of Surveying and Mapping

基　　金：国家自然科学基金资助项目(40271092);香港特别行政区研究基金委员会部分资助项目(编号:香港理工大学5068/99E)

摘　　要：规则格网DEM数据结构简单,易于存储,因而针对于它设计的粗差探测算法成果丰富,较为常用的有基于坡度信息算法以及基于参数统计方法。本文在现有算法的基础上,将主成分分析法应用到DEM粗差定位中,充分考虑DEM数据空间相关的特性,使粗差检测更为准确可靠。本文采用实测的ZX铁路线DEM数据对该算法进行检验,从试验结果可以得知,新算法不仅适用于生产者,也可面向最终用户。Data structure of regular DEM is simple, so it can be stored easily. Correspondingly, the algorithms based on regular DEM are abundant, among which the method based on slope and parametric statistical method had been used frequently. This paper presents a new algorithm for detecting gross errors in DEM by using principal components analysis. The algorithm takes into account the correlated characteristic of DEN, and makes error detection more exactly and credibly. A set of data from ZX railway line is tested using such new algorithm, which proves that the algorithm is not only used for DEM producer, but also for final user. So it has great application area.

关 键 词：数字高程模型 粗差定位 主成分分析法 

分 类 号：P231.5[天文地球—摄影测量与遥感]