消去未知数的分块矩阵混合平差法的原理及其精度评定

本文应用分块矩阵求逆的原理，消去条件方程和混合权函数式中的非观测量未知数，并使条件方程转化为误差方程，这种计算具有明显的规律，且公式化，同时避免了传统间接观测平差法的大量辅助计算工作；混合权函数的转化是利用平差过程中已经算出的有关矩阵来实现的，从而使平差计算更为简捷。

THE PRINCIPLES AND ACCURACY APPRAISAL OF MIXED ADJUSTMENT METHOD BY USING PARTITIONED MATRIX OF ERASED UNKNOWNS

Using the principle of inverse application of partitioned matrix to erass Non-observation unknows in the condition equation and mixed weight function and to convert a condition equation into an error equation, this computation provides an obvious regularity and formulation Thus,a great deal of supplementary computation in the classical approach to the adjustment of indirect observations can be avoided. The transformation of the mixed weight function ie realized by means of the Matrix in the course of the adjustment of values. Hence,the adjustment computation is much simplified and more direct.