Learning Convolutional Sparse Coding on Complex Domain for Interferometric Phase Restoration

Interferometric phase restoration has been investigated for decades and most of the state-of-the-art methods have achieved promising performances InSAR restoration. These generally follow nonlocal filtering processing chain aiming at circumventing staircase effect preserving details variations. In this paper, we propose an alternative approach restoration, i.e. Complex Convolutional Sparse Coding (ComCSC) its gradient regularized version. To our best knowledge, is first time that solve problem in a deconvolutional fashion. The proposed can not only suppress interferometric noise, but also avoid preserve details. Furthermore, they provide insight elementary components phases. experimental results on synthetic realistic high- medium-resolution datasets from TerraSAR-X StripMap Sentinel-1 wide swath mode, respectively, show method outperforms those previous based filters, particularly method: InSAR-BM3D. source code paper will be made publicly available reproducible research inside community.