数字滤波在圆柱体要素提取中的应用

圆柱体截面轮廓要素的测量与提取是圆柱体直径、形位误差和表面粗糙度评价的基础.在介绍2RC滤波、高斯滤波和样条滤波原理的基础上,给出了2RC和高斯滤波的算法的流程图和编制样条滤波程序的思路.对所测10个截面轮廓的要素分别用所编制的2RC、高斯和样条滤波程序依据不同的截止频率(15,50,100和150)进行处理,并对不同滤波方法和不同截止频率下提取要素的质量和速度进行分析,结果表明,在被测截面轮廓要素的滤波方面,样条滤波与2RC和高斯滤波相比,具有明显的优势,2RC滤波存在较明显的非线性相位偏移,运算速度慢且随截止频率变化,而高斯滤波则存在明显的边界效应,且比样条滤波的运行速度慢.

Application of Digital Filtering on the Extraction of Cylindrical Features

The measurement and extraction of the section profile features of cylinders are the basis of the evaluation of their diameters, form & position errors and roughness. On the basis of introducing of the principles of 2RC, Gaussian and spline filtering methods, the flow-plots of 2RC and Gaussian filtering programs and the programming idea of the spline filtering were presented. Ten group of measured section profile features were processed by using three filtering programs above under the different cutoff frequencies, such as 15upr, 50upr, 100upr and 150upr, and the analysis for the quality of the extracted features and the extracting speed were done for the different filtering methods and the different cutoff frequencies. The analysis showed that comparing with 2RC and Gaussian filtering methods, spline filtering method has the advantages in the filtering quality and speed. There exists the nonlinear phase offset for 2RC filtering and its filtering speed is slower than those of Gaussian and spline filtering programs, which is variable for the different cutoff frequencies, while there exists some boundary effective for Gaussian filtering and its filtering speed is slower than that of spline filtering.