Vol. 48, Issue 1, pp. 5-14 (2018)

Abstract

Various methods of deconvolution have been developed for several decades, notably in astronomy and microscopy. The extension of these techniques to the case of a spatially varied blur is currently an open problem. In this work, we consider a zone-invariant point spread function model to take into account blur variation in the image. Thus an algorithm has been used where the minimization of the criterion is performed in parallel on different areas of the image, while taking into account the estimates in the neighboring areas of the sub-images under consideration, so that the final solution is the minimum of the criterion where the blur is spatially varied.