主成分分析在GPS水准数据处理中的应用

Application of Principal Component Analysis in GPS leveling data processing

在GPS水准二次曲面拟合过程中,由于系数阵具有很强的复共线性,使法方程严重病态,在求逆过程中产生过大的扰动误差。为了减小误差影响,本文将主成分分析法运用到了法方程解算中,并通过内符合精度和外符合精度说明该方法的有效性。

In the process of quadric surface fitting of GPS leveling, coefficient matrix has the characteristic of strong multi-collinearity, so the normal equation seems ill conditioned and the perturbation error appears in the process of inverse production. In order to reduce the error influence, Principal Component Analysis （PCA） was applied to the normal equation solution in the paper. The results proved the effectiveness of this method with the inside and outside precisions.