Solar loading infrared thermography method for the non-invasive inspection of movable arts: The contribution of Multivariate Fast Iterative Filtering and Multidimensional Fast Iterative Filtering for 2D data techniques

This work is focalized on the application of two recent techniques, i.e. Multivariate Fast Iterative Filtering (MvFIF) and Multidimensional Fast Iterative Filtering for 2D data (FIF2), that have demonstrated their ability in the pre-processing phase of raw thermal images acquired under the so-called solar loading scenario applied to movable arts (i.e., ancient books and paintings on canvas). The thermographic acquisitions lasted several hours during two separate days. On the one hand, weather forecasts were held in high regard because MvFIF and FIF2 work well in case of signal oscillations caused, in our case, by wind and/or shadows cast. On the other hand, the sun is needed to generate heat transfer through the samples under test (SUT) exposed outdoors. The application of principal component thermography (PCT) to raw and pre-processed thermal images highlights the contribution of MvFIF and FIF2 as a useful step in defect detection. The mathematical part behind FIF2 and MvFIF is underlined in the work, as well as pros and cons of their application in thermographic image processing.