Speckled correlation fringes denoising and demodulation using directional total variation and ridgelet transform

Abstract

One of the challenges in many industrial activities is to analyze the products and investigate their dimensional properties, such as deformations. The digital speckle pattern interferometry technique offers several solutions for the measurement of wide range of parameters (deformations, displacements...) with high accuracy. Generally, there is a correlation between these parameters and the phase of the noisy reflected intensity images (also called correlation fringe patterns or correlograms) of the tested products. Thus, getting access to these parameters requires a good estimation of the phase information. To extract this phase, we propose a ridgelet transform based algorithm for the fringes demodulation, after have been denoised by a new variant of total variation denoising method. These algorithms provide an automatic estimation of the phase feature with high accuracy. Because of such advantages, this method is particularly suitable for real time analyzing of dynamic events even in perturbative environments.