Seasonal Deformation and Accelerated Motion of Infrastructure Monitoring Using a Generalized Differential SAR Tomography

Abstract

Differential synthetic aperture radar tomography (D-TomoSAR) has proven an effective method for retrieving the elevations and deformations of both terrains and urban infrastructures. Particularly, D-TomoSAR with high-resolution X-band SAR data provides even higher accuracy for monitoring of deformations of the man-made structures. However, higher frequency increases the sensitivity of the system to the minute deformations causing a mismatch with the conventional linear deformation model. Currently, the traditional D-TomoSAR system has been extended to higher order versions to extract the seasonal deformation caused by the thermal dilation of structures. Nevertheless, the accelerated motion of scatterers representing another minute deformation component was neglected. As a result, the current D-TomoSAR model cannot accurately retrieve the linear, seasonal, and accelerated motions of the scatterers at the same time. In this letter, a generalized D-TomoSAR system model is proposed by a combination of the polynomial and the seasonal deformation model. Then, an effective algorithm for simultaneously retrieving the multiple deformation components based on the quasi-maximum likelihood (QML) technology is proposed. Finally, numerical simulations and experimental results with real data verify the effectiveness of the proposed system model and algorithm.