摘要
提出了一种新的轮廓滤波器,它以高精度逼近高斯滤波器.这种轮廓滤波器是通过B样条与变分法相结合构造出来.通过调节逼近参数,该滤波器与高斯滤波器的逼近偏差可以控制在0.94%.该滤波器的算法是在频域中实现而不是时域中,所以大大缩减了计算时间.该滤波器的另外一个重要特性是有效地抑制了出现在滤波结果数据边缘的失真,也即边缘效应,这是许多轮廓滤波器所共有的缺陷.对一个实际工件进行测量提取表面中线和粗糙度,结果令人满意.
A low pass profile filter whose transmission characteristics can approximate those of Gaussian filter with high precision has been proposed. This filter was constructed by combining B-spline with variance difference. By regulating the approximating coefficient, the approximation deviation between this filter and Gaussian filter can be shrunken into 0. 94%. The calculation of multiplication was performed in frequency domain rather than in time domain to simplify the methodology and decrease the computation time. Another important property of the filter is that output distortion on the end of data called "end effect" has been effectively restrained, which is the objective for many profile filters. A real complicated engineering surface was used to verify its properties in extraction of mean line and roughness and satisfactory results were obtained.
出处
《纳米技术与精密工程》
EI
CAS
CSCD
2009年第6期490-495,共6页
Nanotechnology and Precision Engineering
基金
国家教育部归国留学人员基金资助项目(AEBQ24403704)