De-noising interferogram inSAR using variance and absolute deviation functions

Abstract

Purpose In this paper, a new efficient method to de-noise the interferometric Synthetic Aperture Radar interferogram, also called wrapped phase image, is proposed with the aim to reduce the residue number and make the phase unwrapping process easy. Design/methodology/approach This method is based on two statistics functions, the former is the phase derivative variance (PDV) defined as a quality map to select the badness areas, the second one is the phase derivative variance (PAD) for a local 3 × 3 pixels filtering which allows to assign an estimated phase for each bad area selected by PDV function. Our filter was tested with a simulated interferograms and compared to other most used filters. Findings With this proposed method, the residues in the interferogram are minimized better than using a conventional filters, and the phase unwrapping process gives a better estimation. Originality/value Combining two statistical functions (PDV and PAD) is efficient in terms of minimizing the noise in the interferogram; this is very helpful to minimize the processing time of the InSAR image particularly the phase unwrapping treatment and have a good quality of the image.