Ground Deformation Retrieval Using Quasi Coherent Targets DInSAR, With Application to Suburban Area of Tianjin, China

Abstract

This paper presents a multiimage DInSAR approach to obtain ground movement using the quasi coherent targets (Q-CTs), such as the bare land and light cultivated areas that distribute widely on the suburban ground surface. The method adopts a new processing strategy for targets selection based on the classification information and coherence properties of the ground objects. Furthermore, the linked coherence model is proposed by defining the parameter of linked coherence on the neighboring Q-CTs network to retrieve the deformation and topographic estimations through multiimage DInSAR analysis. The suburban areas of Tianjin China, with fast township industry development, have been investigated in this study and 21 scenes of ALOS PALSAR dataset, covering the time period August 2007 to November 2010, have been gathered for the algorithm experiment. The experiment results show that there are obvious subsidence patterns detected in this region with subsidence velocity larger than 2 cm/year during the observation period, which have been well validated by both field surveys and the leveling measurements. This study demonstrates the potential of the Q-CTs based DInSAR analysis in the subsidence monitoring of large scale nonurban areas.