Development of image quality metrics for infrared non-destructive testing.

Diverse image processing techniques are used in Infrared for Nondestructive Testing (IRNDT) to enhance images contrast and visibility of cracks or defects in materials. a common procedure to compare the ability of these image processing techniques is subjective. Experts visually compare the outcomes of the techniques and determine which technique, for some given parameters, works better by counting the number of defects made visible in a particular image, however, recently SNR (Signal to Noise Ratio) or detectivity has become a quantitative method for comparison. SNR is an error between a defective area and a sound area and this is not an appropriate mathematical form for image comparison, since image differencing does not adequately capture an estimate of the correlation and contrast between a sound area and a defective area. The main objetive of this proposal is to develop full-reference and non-reference image quality indexes to benchmark IRNDT image processing techniques in the inspection of solar pannels of photovoltaic cells (PV) cells.