Interferometric SAR image restoration using Monte Carlo metropolis method

Andriyan Bayu Suksmono, Akira Hirose

Abstract: We present a novel method of interferometric synthetic aperture radar (InSAR) image restoration. An InSAR image is modeled as a complex-valued Markov random field (CMRF). Corrupted parts, which are indicated by residues in phase data, are restored by using the Monte Carlo Metropolis (MM) method based on their uncorrupted neighbor's CMRF parameter values. The system is implemented as a complex-valued neural network. The restoration process reduces the residue number, which is useful in the phase unwrapping process. The advantage of the method is demonstrated in the unwrapping process of an InSAR image that contains highly dense residues.