Application of double density wavelet transform to surface topographic signal separation

A novel signal separation method based on Double Density Wavelet Transform (DDWT) was presented to reduce the influence of the locations of sampling points on the filtering results from traditional wavelet in surface evaluation. The original profile signal could be decomposed by DDWT into a linear superposition of wavelet functions and scaling functions. After reconstruction of wavelet coefficients corresponding to different components, the required surface topographic discrete signal was obtained. The experiment results show that the separation accuracy of roughness, waveness and other frequency components can be higher about 4% than that of other methods. In engineering surface analysis, the new method can provide near shift-invariance in surface texture separation and extraction, and can decrease the influence of sampling points on filtering results and also improve the measurement precision.