On the Extension of the Minimum Cost Flow Algorithm for Phase Unwrapping of Multitemporal Differential SAR Interferograms

In this paper, an extension of the minimum cost flow (MCF) algorithm dealing with a sparse data grid, which allows the unwrapping of multitemporal differential synthetic aperture radar (SAR) interferograms for the generation of deformation time series, is presented. The proposed approach exploits both the spatial characteristics and the temporal relationships among multiple interferograms relevant to a properly chosen sequence. In particular, the presented solution involves two main steps: first of all, for each arc connecting neighboring pixels on the interferometric azimuth/range grid, the unwrapped phase gradients are estimated via the MCF technique applied in the temporal/perpendicular baseline plane. Following this step, these estimates are used as a starting point for the spatial-unwrapping operation implemented again via the MCF approach but carried out in the azimuth/range plane. The presented results, achieved on simulated and real European Remote Sensing satellite SAR data, confirm the effectiveness of the extended MCF unwrapping algorithm