Update two-dimensional SAR offset tracking deformation time series with complex sequential least squares estimation

Abstract

ABSTRACT Unlike synthetic aperture radar (SAR) interferometry, SAR offset tracking method can measure two-dimensional (2-D) large-gradient deformation with precision as 1/10 to 1/30 pixels size. In terms of 2-D deformation time series estimation, the pixel-offset small baseline subset (PO-SBAS) method was proposed. However, little research has been conducted on the dynamic estimation of 2-D deformation time series for new SAR acquisitions, which has increasing demand with the availability of SAR data. In this letter, we introduced a novel method to update 2-D deformation time series dynamically under the frame of PO-SBAS. Firstly, complex least squares (CLS) method is derived to solve the complex linear function model, and the difference between the CLS estimation in complex value and the LS estimation in real value for the real and imaginary parts, respectively, are compared. Secondly, complex sequential least squares (CSLS) estimation is proposed to update 2-D deformation time series dynamically and efficiently. Finally, both simulated and real SAR data verified the performance of the proposed method. It can be taken as a promising SAR tool for dynamic 2-D deformation estimation for the scenarios of glacier movement, mining-induced displacement, and co-seismic deformation, etc.