InSAR phase analysis: Phase unwrapping for noisy SAR interferograms

Abstract

´ ´ In this chapter, we proposed a modified filtering algorithm to the Lopez and Fabregas (LopezMart´nez & Fabregas, 2002) noise reduction algorithm for the interferometric phase noise in i SAR interferometry using a multiresolution approach. Our contribution to the existing algorithm consists on the exploitation of the InSAR coherence map in order to generate a more adaptive mask for each decomposition level. Moreover, we presented a general probabilistic framework to phase unwrapping problem based on the work of Rodriguez and Servin (Rodriguez-Vera & Servin, 1995). Both MRF and Bayesian estimator were applied to recover the desired phase, as the optimal field solution of the maximum a posteriori (MAP) estimation criterion. An iterative method that minimizes a general energy function is proposed, and a parallel algorithm based on gradient descent optimization is designed to perform this task. The proposed solution overcomes some important limitations of most of the phase unwrapping procedures, and the results show robustness, and stability. These results give us new ideas for the applications of the InSAR unwrapping phase MRF algorithm for unwrapping interferometric synthetic Aperture Sonar (InSAS) (Bonifant et al., 2000) phase. Both two phases present similar statistics. However, if the ground truth could be used for the InSAR result validations, it not the case for the InSAS results. In fact, the ability to produce full coverage bathymetric maps and generate accurate measurements of the seafloor height, is limited. So, an adaptive unwrapping procedure could be interesting.