大规模联合平差中的方差分量估计并行设计

在采用方差分量估计消除大规模联合平差系统随机特性差异时,需要进行大量大型法方程相关计算,往往使得到随机模型的效率非常低。本文基于并行设计的思想,结合大型矩阵的相关高效计算方法,提出在大规模联合平差中进行并行方差分量估计的基本流程,理论上能显著提高估计效率。

Parallel Design of Variance Component Estimation in Large Scale Combined Adjustment

When the variance component estimation is used to eliminate the discrepancy of the stochastic characteristics of a large-scale combined adjustment system,a large number of large scale equation related calculations are needed,which often makes the efficiency of the stochastic model very low.In this paper,based on the idea of parallel design,the basic flow of the parallel variance component estimation in large-scale combined adjustment is put forward in combination with the high efficient calculation method of large matrix.In theory,the estimation efficiency can be greatly improved.