| Summary: | Synthetic Aperture Radar (SAR) has the capability to map the Earth with high resolution and is used in both civilian and military applications. The development of SAR systems is going towards higher resolution. Today there are Wide Band (WB) SAR instruments that have a resolution actually smaller than the operating wavelength. One example is the CARABAS sensor developed by the Swedish Defence Research Establishment (FOA). It operates at VHF and has shown many unique capabilities such as the detection of concealed targets. Great efforts are being made to build WB SAR systems at higher frequencies. One of these projects is the LORA project at FOA. The LORA system will operate on UHF and it will have several antenna channels which can be used for space-time processing, e.g. to suppress clutter and jammers, and to detect ground moving targets. In this thesis we have considered two different areas connected with SAR. In the first part we develop space-time adaptive algorithms for WB SAR systems with array antennas. The aim of the work is to detect ground moving targets and to suppress strong clutter and jammer noise. A likelihood ratio test is proposed for detection. The method combines fast backprojection algorithms with target detection. Moving targets are focused using their proper relative speed. The methodology has been tested on narrow band data. Bistatic antenna effects on WB SAR systems for moving target indication are discussed. Simulations of WB SAR using combined fast backprojection and clutter suppression are presented. Jammer suppression is tested for WB SAR, using space-time adaptive processing. The results are from an experiment with the CARABAS radar. In the second part we examine what a microwave radar sensor actually senses in sea ice and we investigate the capability of the SAR onboard the ERS-1 satellite to classify Arctic sea ice. Polarimetric scatterometer measurements have been carried out at microwave frequencies (L-, S-, C- and X-band). The measurements are compared with scattering models. ...
|
|---|