三维坐标转换的多元线性回归模型及粗差剔除

The multiple linear regression model of 3D coordinates transformation and gross errors detection method

将三维直角坐标转换归结为多元线性回归模型，导出其矩阵方程形式及精度评定公式；通过实例研究多元线性回归和7参数解法的异同，得出重要结论；提出一种构造τ统计量逐点剔除观测值粗差的有效方法。

It provides the multiple linear regression model for a solution to the 3D coordinates transformation, developing a matrix equation and a set of formulas for precision evaluation. It draws a great important conclusion through the similarities and udifferences between the multiple linear regression model and the 7-paramater model supported by some cases. Furthermore, it explores an effective way of constructing τ statistics to eliminate gross errors from the surveying data.