A Space-Time Extension of the Minimum Cost Flow Phase Unwrapping Algorithm for DInSAR Time-Series Generation

Abstract:
We propose in this work an extension of the Phase Unwrapping (PhU) minimum cost flow (MCF) algorithm that exploits both the spatial properties and the temporal relationships among a set of multiple Differential Synthetic Aperture Radar (DInSAR) interferograms, relevant to a properly chosen sequence of SAR images. The proposed approach relies on the use of two distinctive Delaunay triangulations: the former is relevant to the SAR data distribution in the “Temporal/Perpendicular baseline” plane, the latter is carried out in the Azimuth/Range plane and involves the coherent pixels common to the generated interferograms. In particular, the unwrapping operation of the overall data set is performed via a two-step processing procedure carried out by applying the basic minimum cost flow approach to the two different triangulations. In order to validate the algorithm, a number of experiments on real SAR data were accomplished; the obtained results demonstrate the effectiveness of the proposed PhU approach.