抗差最小二乘配置在数字化地图几何纠正中的应用讨论.pdfVIP

第 31卷 第 2期 测绘与空间地理信息 Vol. 31, No. 2 2008年 4月 GE OMA TICS SPA TIAL IN FORMA TION TECHN OLOGY Apr., 2008 张 亮 (长安大学 地质工程与测绘学院, 陕西 西安 710054) :抗差估计具有较好的抗拒异常观测值及粗差的能力, 而最 二乘配置又能较好地处理系统误差, 本文结 合两者的优点, 利用抗差最 二乘配置对数字化地图进行几何纠正, 其中对协方差函数采用抗差拟合,得到了较 好的结果实验证明在 GIS数据处理的扫描数字化地图几何纠正中, 抗差最 二乘配置在抗拒异常值和处理系 统误差方面优于单纯的最 二乘估计和单纯的最 二乘配置方法 :抗差; 最 二乘配置; 最 二乘估计; 协方差函数 : TP391. 72 : B : 1672- 5867(2008)02- 0145- 04 Discussion on theApplication ofRobustLeast Squares Collocation inGeo etricCorrection ofDigitalMap Z ANG Liang (College ofGeology Engineering andGeo atics, ChanganUniversity, Xican710054, China) Abstract: Robust estmi ation can resist abnormal observations and coarse error wel,l and least squares collocation can preferably process system ic errors. Combined with their respective advantages, this paper mi plemented the geometric correction for digitalmap u- sing robust least squares collocation, and obtained satisfied results. The procedure in robust fitting for covariance function is also dis- cussed. The expermi entalresults showed that the robust least squares collocation is better than pure least squares and least squares col- location in resisting abnormalobservation and processing systemic errors in the geometric correction ofscanned digitalmap forGIS data process. Key words: robust estmi ation; least squares collocation; least squares estmi ation; covariance function , 0 , , , () , : 1 , ; , : , u, 1 L = B X + A