A Phase Unwrapping Approach Based on Extended Kalman Filter for Subsidence Monitoring Using Persistent Scatterer Time Series Interferometry

Abstract

Persistent scatterer interferometry that overcomes the spatial and temporal problems is an effective technique for measuring the surface deformation. The main issue of the technique, however, is unwrapping the persistent scatterers’ phases that are irregularly spaced. The problem is even more challenging when the deformation rate is too high. In this paper, a phase unwrapping method based on an extended Kalman filter as a path-following method is presented. Mean fringe frequency as the phase slope, which is to be used in the extended Kalman filter, is estimated within a local window. The most proper paths for unwrapping are identified by a proposed hybrid cost map, which is a combination of the scaled phase derivative variance, the distances between the persistent scatterer pixels, and the discontinuity locations. The presented method is applied to ENVISAT ASAR images acquired over southwestern Tehran for subsidence monitoring. The maximum displacement rate estimated from the proposed method is about 18 cm/year. Moreover, the results are then validated using a continuous GPS station and leveling measurements acquired over a different time interval. The difference between the rates estimated from the GPS measurements and the proposed method is 0.2 cm/year. It was found that the proposed method makes a significant improvement over the unwrapping problem.