Spatial-Temporal Phase Unwrapping Method for GB-InSAR Based on Kalman Filter

Ground-based interferometric synthetic aperture radar (GB-InSAR) can effectively achieve high-accuracy measurement of surface deformation, which is widely used in the field of landslide monitoring and warning. The accuracy of phase unwrapping (PU) is one of the key factors affecting the accuracy of deformation measurement. This paper proposes a spatial-temporal PU method based on the Kalman filter to address the errors of PU caused by large interferometric phase gradients between deformation pixels and stable pixels. Firstly, for stable pixels, the minimum cost flow method is used for PU. Then, for the deformation pixels, the temporal domain Kalman filter is employed to estimate the unwrapped phase and obtain the estimation of ambiguity number in the spatial domain. Finally, optimizing the objective function of PU by minimizing the estimation error of ambiguity number improves the unwrapping accuracy for the deformation pixels. The results of simulation and measured data have shown that this method can effectively correct the phase unwrapping error caused by large interferometric phase gradients and improve the reliability of deformation inversion.