High-accuracy Digital Elkevation Model Generation ans ship monitoring from Synthetic Aperture Radar images: innovative techniques and experimental results

In this Thesis several state-of-the-art and innovative techniques for Digital Elevation Model (DEM) generation from Synthetic Aperture Radar (SAR) images are deeply analyzed, with a special focus on the methods which allow the improvement of the accuracy of the DEM product, which is directly related to the geolocation accuracy of geocoded images and is considered as an enabling factor for a large series of civilian and Defence applications. Furthermore, some of the proposed techniques, which are based both on phase and amplitude information, are experimented on real data, i.e. COSMO-SkyMed (CSK) data, assessing the achievable performances compared with the state-of-the- art, and pointing out and quantitatively highlighting the acquisition and processing strategies which would allow to maximize the quality of the results. Moreover, a critical analysis is performed about the main errors affecting the applied techniques, as well as the limitations of the orbital configurations, identifying several complementary techniques which would allow to overcome or mitigate the observed drawbacks. An innovative procedure for on-demand DEM production from CSK SAR data is elaborated and proposed, as well as an auto-validation technique which would enable the validation of the produced DEM also where vertical ground truths are not available. Based on the obtained results and on the consequent critical analysis, several interferometric specifications for new generation SAR satellites are identified. Finally, a literature review is proposed about the main state-of-the-art ship monitoring techniques, considered as one of the main fields of application which takes benefit from SAR data, based on single/multi-platform multi-channel SAR data, with a focus on TanDEM-X (TDX).