Automatic frames extraction and visualization from noisy fringe sequences for data recovering in a portable digital speckle pattern interferometer for NDI

We report on a method for recovering data from a simple portable Digital Speckle Pattern Interferometer, we used for monitoring structural behavior of a painting on wood, hanging on a wall, outside of laboratory conditions, without anti-vibration devices. In such a situation, fringes produced by the object displacements were affected by unpredictable distortions caused by environment vibrations. However, an sufficient number of suitable, i.e., undistorted or barely distorted, fringe patterns usable for processing was found. We performed multiple acquisitions at a frame rate as high as possible. The main task was picking out usable interferograms from large amount of frames. We developed a PC-based method based on jointly analysis of spectral content and fringe image sharpness as selection rules. The selected frames were utilized for off-line processing by using an approach based on Hilbert Transform and Phase Unwrapping via MAx-flow (PUMA) algorithm. We obtained a collection of displacement-maps, that allowed for evaluating the whole structure deformations, caused by environmental thermo-hygrometric fluctuations.