Gaussian filtering algorithm describing the topography of temper rolled strip and related edge effect

Abstract

Abstract The surface topography of temper rolled strip consists of a range of spatial wavelengths with approximate normal distribution in height. Gaussian filtering methods can be applied in measuring and characterizing the topography of temper rolled strip. In this study, three Gaussian filtering algorithms, namely convolution algorithm, fast Fourier transform algorithm and fast recursive algorithm, are compared, in order to find out the most suitable one for the topography of temper rolled strip. A similar profile extension method based on the distribution of the topography of strip is proposed to eliminate the edge effect. The results show that the convolution algorithm is obviously beneficial for the computational efficiency. It is also found that the similar profile extension method can effectively eliminate the edge effect and improve the reliability of edge data. The Gaussian convolution algorithm, combined with the similar profile extension method, is able to effectively assess the surface topography of temper rolled strip.