DEVELOPMENTS AND EXPERIMENTS OF MULTIDIMENSIONAL DIFFERENTIAL SAR TOMOGRAPHY FOR SATELLITE URBAN MONITORING

After the maturation of the interferometric SAR (InSAR, D-InSAR) techniques based on phase-only data, much interest is growing in techniques based on the coherent combination of complex (i.e. amplitude and phase) SAR data for the extraction of even more rich information on the observed scene, to fully exploit existing SAR data archives, experimented multi-antenna airborne and unmanned air vehicle systems and new multistatic satellite clusters. Among these techniques, 3D SAR Tomography (Tomo-SAR) is an experimental multibaseline (MB) interferometric mode achieving full 3D imaging in the range-azimuth-height space through elevation beamforming, i.e. spatial (baseline) spectral estimation. Tomo-SAR resolves multiple scatterers in height in the same cell (layover effect), overcoming a limitation of conventional InSAR processing and allowing the analysis of complex scenarios. Recently, a more advanced coherent data combination mode termed Differential SAR Tomography (Diff-Tomo) has been originated at University of Pisa, synergically integrating the D-InSAR and the Tomo-SAR concepts to allow “opening” the SAR pixel in complex non-stationary scenes, degarbling its information content. Diff-Tomo is based on two-dimensional space-time spectral analysis, fully exploiting the MB-multipass data. It allows the joint resolution of multiple heights and deformation velocities of the scatterers mapped in a SAR pixel. All these methods can be set in the more general framework of computational imaging, where computation plays an integral role in the image formation process. In this work, new processing methods and aspects of this new framework are investigated for complex scattering urban scenarios. The first analyzed aspect is the possibility to extend advanced two-dimensional space-time spectral analysis methods (Adaptive Beamforming and Model-based methods), which require spatial multilooking (i.e. loss of range-azimuth resolution), to single-look processing also maintaining a low computational burden. In fact, especially for urban scenarios, single-look processing is preferable to preserve the full horizontal resolution, that is of meter order in the new generation X-band SAR systems (TerraSAR-X, Cosmo-SkyMed). The second analyzed aspect regard the possibility to estimate non-linear motion parameters from the data, maintaining the full horizontal resolution. Then, it has been introduced and tested a unique “5D” single-look Diff-Tomo method to estimate different seasonal thermal dilation motions of multiple scatterers. These developments of Multidimentional SAR imaging are discussed through extensive simulated analyses and ERS 1/2 and Cosmo-SkyMed data results.