A Tomographic Approach to Multi-Pass SAR Imaging
Seconda Università di Napoli,
Via Roma 29,
(2) Universita Mediterranea di Reggio Calabria, Via Graziella, I-891000, Italy
In this communication we propose the employment of a tomographic technique in the framework of SAR imaging. In particular, for a two-dimensional and scalar geometry we cast the imaging problem as the inversion of the linearized scattering operator. In this way the imaging is achieved discarding the usual techniques of range compression and azimuth focusing. We consider a multi-frequency and multi- monostatic configuration in which the different point of illumination and observation arise from different passes of the SAR sensor. Furthermore, the scenario under test is supposed time- invariant so no de-correlation occurs for the different observations. We also suppose to a priori know that the scattering object is enclosed in square investigation domain and perform the analysis and the inversion by through the Singular Value Decomposition (in short SVD) of the relevant operator. This allows to determine an effective base functions in order to represent in a non- redundant way the unknown scattering object and thus to counteract the instability of reconstructions due to the effect of noise. Moreover, a sampling criterion of the scattered field, that is the number and the positions the sensor has to occupy, and a resolution estimate achievable in the reconstructions, are derived. Finally, we test the proposed approach on a case study by showing the capability of the scheme to localize and reconstruct objects located at different “height” without exploiting interferometric information.