表面粗糙度测量中的高斯滤波快速算法

应用高斯函数的逼近法和冲激响应不变法，设计出用于表面粗糙度测量的高斯数字滤波器，并给出了其零相移的递归滤波算法，算法简洁，易于实现。递归计算方法的计算量小，计算效率高。适当增加滤波器节数，在不增加很多计算量的情况下，可以达到很高的精度。文中给出的滤波器例子，传输偏差大约为2％和1％，处理一次表面测量轮廓数据，在当今最普通的计算机上，滤波时间小于1s。

A Fast Gaussian Filtering Algorithm for Surface Roughness Measurements

The Gaussian filters for surface roughness measurements are designed by using both the Gaussian function's approximation method and the impulse invariant transformation. The zero-phase-shift recursive algorithms of the Gaussian approximation filters are so simple that they are completed easily. There are a small number of computation operations in these recursive algorithms, the filtering algorithms are therefore efficient. The Gaussian approximation filters, consisting of more cascaded basic filters, result in higher accuracy of the amplitude transmission characteristics without increasing a large number of computation operations. Two filtering algorithms, with less then 2% and 1% of the transmission characteristic deviations and less then 1 second needed for a 2D roughness filtering using the currently common computers, are given.