うねり曲線を抽出するローパスフィルタの研究(第2報) : エンド効果の低減

概要

論文の詳細を見る
A Gaussian filter is widely used as a low-pass filter for surface roughness measurement. It has two disadvantage : a high calculation cost and an end effect. To eliminate these disadvantages, a spline filter is proposed. However, due to its greater profile error, it cannot replace a Gaussian filter. In a previous report, the authors proposed a method that allows a Gaussian filter to be replaced with a filter that offers the advantages of low calculation cost and very small profile error. More specifically, this method employs a down-sampling process and a G-spline filter. However, it has a disadvantage, that is, the end effect cannot be reduced. In this paper, we report a method of reducing the end effect by replacing the data string end with data obtained from a regression line, followed by pre-processing based on point symmetric extension and shearing, and finally by post-processing (i.e., an inverse transformation process of pre-processing) after down sampling.
社団法人精密工学会の論文
2005-06-05