总体最小二乘法在坐标转换中的应用

Application of Total Square in Coordinate Transformation

在处理坐标转换数据的方法中,通常使用的方法是最小二乘法,但其由于不能顾及系数矩阵误差而具有一定的局限性,导致坐标转换结果的可靠性较差。因此,需要一种新的方法来弥补最小二乘法的不足。本文引入总体最小二乘法和混合最小二乘法,采用仿真数据求解坐标转换七参数,并将结果与其仿真值进行比较,证明采用混合最小二乘法得到的坐标转换七参数更接近于理论值。

In the method of processing coordinate transformation data,the usual method is least square method,but because it cannot take account the coefficient matrix error and has certain limitation,the reliability of the coordinate transformation results is relatively poor.Therefore,a new method is needed to compensate for the deficiency of least square method.The total least square method and total least square-least square method was induced into this paper,adopt simulation data to calculate seven parameters of coordinate transformation,compare the results with its simulation values,It is proved that the coordinate transformation seven parameters obtained by the hybrid least square method is closer to the theoretical value.