数字化曲线的整体最小二乘平差

Total least squares method for digitized curve fitting

针对一般数字化曲线拟合方法的不足,提出了利用整体最小二乘平差计算数字化曲线的参数。本文以直线拟合为例,在参数平差中,将数字化曲线的X坐标和Y坐标均看作观测值,采用整体最小二乘平差法进行平差和精度评定。算例结果说明整体最小二乘平差法拟合直线方程精度更高。

For improving shortcoming of fitting methods for the digitized curve, this paper applies the total least squares adjustment method to calculate parameters of digitized curves. To the line fitting as an example, X and Y coordinates of the digitized curve are observations, and using the total least squares method to adjust and evaluate accuracy of parameters. The calculation results show that the method has the capability to increase the accuracy of parameters estimated.