方差分量估计的MINQE理论及其在摄影测量中的应用

本文深入研究了异方差分量模型的MINQE(U,I)准则的算法实现的可能性,所导出的约化算法具有压缩内存,提高运算速度等明显优点,因而具有实用意义。文中介绍了实现约化算法的计算机程序框图,并提供了数据例子。作者将方差分量估计的MINQE理论引入解析摄影测量平差实践,分别采用模拟数据和新滩滑坡监测数据研究了自检校光束法平差中随机模型的确定,结果表明,MINQE理论用于自检校的验后权估计是可行的。最后,作者就平差随机模型的扩展,以及MINQE理论的应用提出了进一步研究的几点设想。

The MINQE Theory of Variance Component Estimation and Its Applications in Photogrammetry

This paper derives a reduced algorithm of MINQE(U,I) criterion and investigates thepossibility of applying it in heteroscedastic variance component models. The obviousadvantages of this algorithm are compressed computer memory and the speeding up ofoperations. A computer program block diagram and a numerical example are given. Theauthcr of the paper applies the MINQE theory to self-calibration bundle adjustment inphotogrammetry, and to computations of monitoring deformations of the XINTAN landslide.The result shows a possibility to use the MINQE theory for improving adjustment randommodel. Several suggestions regarding random models and the MINQE theory to be dis-cussed further are proposed.