Parametric differential SAR tomography of decorrelating volume scatterers

Much interest is continuing to grow in advanced interferometric SAR methods for full 3D imaging, in particular of volumetric forest scatterers. Multibaseline SAR tomographic elevation beam forming i.e. spatial spectral estimation is a promising technique in this framework. In this paper, the important effect of temporal decorrelation during the repeat pass multibaseline acquisition is analyzed on superresolution (MUSIC) tomography with limited sparse data. Moreover, the ultimate achievable estimation performance of the volume height parameters is investigated by deriving the Cramer-Rao bound in the partially coherent scenario. Finally, new tomographic methods are proposed robust to temporal decorrelation phenomena, exploiting the advanced differential tomography concept which produces `space-time´ signatures of scattering dynamics in the SAR cell. To this aim, the 2D version of MUSIC, and a generalized MUSIC matched to non-line spectra, are applied to decouple the nuisance temporal signal histories in the spatial spectral estimation. First differential tomography results with real P-band forest data are presented.