Study of various preprocessing schemes and wavelet filters for speckle noise reduction in digital speckle pattern interferometric fringes

Abstract

A new filtering scheme is investigated for the removal of speckle noise in the recorded fringe pattern by a digital speckle pattern interferometric technique (DSPI). The scheme consists of a preprocessing scheme i.e., averaging, sampling, thresholding, and again averaging, followed either by a symlet or biorthogonal wavelet filter. The preprocessing scheme improves the contrast between the dark and bright fringes, and the implementation of the biorthogonal wavelet filter, which is symmetrical and purely linear in phase, solves the problem of border effect and discontinuity. The developed scheme is implemented on recorded fringe patterns of a vibrating cantilever beam and vibrating square plate with all ends fixed. The performance and effectiveness of the proposed scheme in reducing speckle noise is evaluated and compared in terms of reduction in speckle index and the improvement of the signal-to-noise ratio (SNR) of speckle interferograms. The new scheme has the ability to improve the speckle index and SNR significantly.