针对光学线阵遥感影像几何定位中, 系统误差改正参数对应系数矩阵存在误差的问题, 提出了一种基于正则化总体最小二乘的光学线阵遥感影像定位方法。首先依据有理函数模型的定义, 构建共线条件方程, 利用线性化构建光学线阵影像定位方法和系统误差改正方法, 然后依据EIV模型的定义和性质构建相应的优化目标函数, 并引入正则化项, 依据Lagrange条件极值原理推导基于正则化总体最小二乘的系统误差参数迭代估计方法。实验结果表明: 与经典的最小二乘平差算法相比, 该方法的总体定位精度提高了11.61%, 且比Tikhonov正则化法的定位精度平均提高了6.06%。本文提出方法在不增加任何额外控制信息的情况下, 是提高光学线阵影像定位精度的有效途径。

In geometric positioning of optical line array remote sensing image, as it exists error on system error correction parameters corresponding coefficient matrix, a geo-positioning method of remote sensing image of optical line array based on regularized total least squares was proposed in this paper. Firstly, the collinear equation was constructed based on the definition of rational function model, Secondly, the geo-positioning and systematic error correction methods were derived by linearization. Then, the corresponding optimal objective function was constructed based on the definition of the EIV model with a regularization term introduced. The iterative solution method of system error parameters based on regularized total least squares solver was deduced by the Lagrange conditional extremum. Experiment results indicate that, the geo-positioning accuracy of this method is increased by 11.61% compared with the classical least squares adjustment algorithm, and it is increased by 6.06% in contrast to Tikhonov regularization geo-positioning accuracy. The proposed method in this paper can increase the geo-positioning accuracy of linear array CCD images effectively without any extra information.