Dynamic Estimation of Multi-Dimensional Deformation Time Series From InSAR Based on Kalman Filter and Strain Model

Jihong Liu,Jun Hu,Zhiwei Li,Qian Sun,Zhangfeng Ma,Jianjun Zhu,Yaxin Wen

DOI: https://doi.org/10.1109/tgrs.2021.3125574

IF: 8.2

2022-01-01

IEEE Transactions on Geoscience and Remote Sensing

Abstract:With the increasing amount of synthetic aperture radar (SAR) data with various imaging geometries (at least ascending/descending tracks), it is possible to obtain accurate multi-dimensional (MD) deformation time series with long time span. However, in most cases SAR data of different geometries are un-synchronously acquired over the same region, making it impossible to directly solve the underdetermined observation model (OSM) between the interferometric SAR (InSAR) measurements and the MD deformations. Kalman filter (KF), as one of the most famous dynamic estimators, can obtain a priori information of the unknowns based on the preexisting time series, therefore it can be used to deal with this InSAR underdetermined problem. This article employs the KF to realize the dynamic estimation of MD deformations with short-baseline interferograms. The innovation lies in the establishment of the KF state transition model (STM) and OSM, which aims to make the InSAR monitoring problem better adapt to the KF. Particularly, by assuming a smooth deforming process, existing deformation time series are used to establish the STM and to predict the deformations at current moment. Besides, a strain model (SM) is employed to assist the establishment of the OSM. Simulation and real experiments in the Geysers geothermal field (GGF), U.S. demonstrate that, compared with the state-of-the-art methods, the proposed KF method allows more robust deformation estimation and achieves higher computational efficiency for dynamic estimation.