Phase Unwrapping with Quality Map and Sparse-Inpainting in Interferometric SAR

Inhalt:
Phase unwrapping is an important step to construct Digital Elevation Model (DEM) in Interferometric Synthetic Aperture Radar (IFSAR). Phase unwrapping is the reconstruction process of the true phase from the received phase which lies in the interval [-π,π]. In this paper, a global optimization approach is used based on the first order Markov Random Field (MRF) coefficients. By considering different quality of the received phases, we apply the new quality map-related coefficients in order to introduce a new energy function. With minimizing the new energy function, phase unwrapping is done. To improve the quality of unwrapped phase, applying the prior information about the smoothness feature of the true phase, and with using the characteristic of the derivative operator which identifies variations, we estimate and replace pixel values with high variation based on sparseinpainting method. Experimental results explicitly confirm the efficiency of the proposed method.