Array Signal Processing for Synthetic Aperture Radar (SAR)

Synthetic aperture radar (SAR) is a kind of imaging radar that can
produce high resolution images of targets and terrain on the ground.
At present, most of SAR processing algorithms are based on matched
filtering. This method is easy to implement and can produce stable
results. However, It also has some limitations. This approach must
obey the Nyquist sampling theorem and the resolution depends on
bandwidth of pulses. This means that the matched filter approach
must be based on a large amount of raw data but the performance is
limited. With the development of radar imaging, it is difficult for the
matched filtering approach to meet the requirement of high resolution
SAR images.
In this thesis, a new processing method based on the least squares
(LS) beamforming is utilized in the processing of SAR raw data. The
model of SAR simulates a virtual linear array. The processing of
SAR data can also be seen as a process of beamforming. The 1-
D azimuth direction echo data is processed using the beamforming
method. Simulation results based on the least squares design method
are compared with the matched filtering method and the conventional
beamforming method with different windows.