Hyperimage concept: Multidimensional Time-Frequency Analysis applied to SAR imaging

This paper deals with the analysis of non-stationary scatterers in SAR images. Indeed, SAR imaging makes the assumptions that the scatterers are isotropic and white in the emitted frequency band. However, new SAR applications use a large bandwidth and a strong angular excursion. These assumptions become obsolete and the behavior of scatterers becomes non-stationary. The basic tool to study non-stationary signals is the time-frequency analysis. Recent studies based on multidimensional Time-Frequency Analysis describing the angular and frequency behavior of scatterers has highlighted anisotropic and dispersive behavior of bright points. This paper generalizes the hyperimage concept to study scatterers. Multidimensional Time-Frequency distributions are tested on simulations, then they are applied to very high resolution SAR images and show some scatterers are anisotropic and dispersive.