| 基于总体最小二乘的直线拟合方法探究 |
| |
| 引用本文： | 鞠英博,李伟,姚磊.基于总体最小二乘的直线拟合方法探究[J].测绘与空间地理信息,2017(6):166-168. |
| |
| 作者姓名： | 鞠英博 李伟 姚磊 |
| |
| 作者单位： | 1. 黑龙江第二测绘工程院,黑龙江哈尔滨,150025;2. 上海市测绘院,上海,200129 |
| |
| 摘 要： | 在直线拟合问题中,经典的最小二乘拟合方法在自变量选取不同时,拟合的参数值和中误差存在较大差别,故本文利用模拟数据对经典最小二乘和总体最小二乘拟合结果进行对比分析,得出结论认为:经典最小二乘自变量选取不同结算参数的原因是在进行拟合计算时忽略了自变量的误差,使拟合结果只能在一个方向上保持最佳;利用总体最小二乘参数拟合的方法进行直线拟合时拟合结果不受自变量变化的影响,并能够提高拟合精度。
|
| 关 键 词： | 最小二乘法 总体最小二乘 直线拟合 EVI模型 SVD分解 |
Linear Fitting Method Based on Total Least Squares |
| |
| JU Ying-bo,LI Wei,YAO Lei.Linear Fitting Method Based on Total Least Squares[J].Geomatics & Spatial Information Technology,2017(6):166-168. |
| |
| Authors: | JU Ying-bo LI Wei YAO Lei |
| |
| Abstract: | In the linear fitting problem,because the classical least squares fitting method is not the same as the independent variables,the fitting parameters and the errors have great difference,in this paper,the simulation results are used to compare the classical least squares and total least squares fitting results:The reason for the different parameters of the classical least squares independent variables is that the errors in the independent variables are ignored in the fitting calculation,the fitting result can only be kept in one direction;By using the method of least square parameter fitting,the fitting result is not affected by the independent variable and improve the fitting precision. |
| |
| Keywords: | least square method total least squares straight line fitting EVI model SVD decomposition |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
