|Thesis abstract: |
The impact of human activity on the global environment has dramatically increased in the last two centuries. The necessity of understanding the effect of the human development on the natural equilibrium led to many efforts in this direction. On the one hand measures of Earth parameters multiplied; on the other hand, modeling attempts to describe the Earth dynamics relying on such measurements spread. Of particular interest in this scenario is the study of the above the ground biomass together with its the relationship with carbon cycle and global warming. Synthetic Aperture Radars (SARs) working at microwaves offer a perfect tool for global mapping due to its capability of imaging day and night, with any atmospheric condition and due to the fast coverage of very large areas. In particular longer wavelength SARs enable to penetrate the vegetation layer up to the ground level. It follows that each depth inside of the vegetation and the underlying ground contribute to the radar signal. The whole forest layer can by imaged, still refined processing chains are needed to interpret this data properly. SAR tomography has been widely exploited throughout this work as the main operational tool able to separate contributions coming from different heights inside of the vegetation layer. SAR tomography is able to retrieve an image of the whole vegetation layer without a priori assumptions on the target; results coming from SAR tomography can be considered as the starting point for the subsequent modeling. This thesis is focused on experiments on forested areas with the aim of offering a clearer view of the interaction of the electromagnetic wave with such a complicated target. Different processing chains on different dataset are here shown to highlight different features of the ground under the forest offered by SAR imaging. The reliability of physical parameter is subjected to a good calibration of the data, that is to a minimization of the impact of the imaging system in favor of the imaged target. For this reason the issue of calibrating SAR data coming from natural targets is covered almost in each chapter; a novel calibration algorithm on forests is here presented too.