Cross Sensor Simulation of Tomographic SAR Stacks

Abstract

This paper presents an algorithm for simulating tomographic synthetic aperture radar (SAR) data based on another stack actually gathered by a real acquisition system. Through the procedure here proposed, the simulated system can be evaluated according to its capability to image complex natural media rather than reference point targets. This feature is particularly important whenever the biophysical properties of the target of interest must be preserved and cannot be easily modeled. The system to be simulated may be different from the original one concerning resolution, off-nadir angles, bandwidth and central frequency. The algorithm here proposed handles these differences by properly taking into account the wavenumbers of the target illuminated by the real survey and requested by the simulated one. The complex images constituting the synthetic stack are associated with the effective vertical interferometric wavenumber peculiar of the geometry to be simulated, regardless of the original data. Furthermore, the three-dimensional resolution cell of the simulated tomographic system is consistent with the simulated geometry concerning size and spatial orientation. These two latter features cannot be guaranteed by simply filtering the original stack. The simulator here proposed has been used to simulate the tomographic stack expected from the forthcoming European Space Agency (ESA) BIOMASS mission. The relationship between baseline distribution and 3D focusing capability was explored; special attention has been paid to the robustness of tomographic power at being a good proxy for the above ground biomass in tropical regions.

Ground and Volume Decomposition as a Proxy for AGB from P-Band SAR Data

“…TomoSAR will be one the main innovations characterizing the BIOMASS mission, allowing the ability to single out returns from different layers of the forest canopy. The BIOMASS tomographic phase [18] will provide the first global coverage during the first 14 months of operations, by collecting seven ascending and seven descending passes with 3-days repeat pass time, designed to maintain coherency at P-band [19]. Thanks to the long wavelength of about 70 cm it will be possible not only to characterize electromagnetic interactions of radar waves with the forest structure and generate AGB maps over tropical biomes, but also to generate a global map of terrain topography below the vegetation layer (DTM) [17] and subsurface features [20,21].…”