Robust processing of SAR hologram data to mitigate impulse noise impairments

The standard linear algorithm for synthetic aperture radar (SAR) image synthesis consists of fusing the real and imaginary components of the hologram data, and applying a matched filter to the radar returns from each point to generate the SAR image. This algorithm becomes inefficient when the SAR hologram data are corrupted by impulse noise due to data transmission and coding/decoding errors. The degradation of synthesized SAR images in the presence of impulse noise in the hologram data is considered and new nonlinear algorithms based on robust estimation are proposed. Experimental results for measured SAR data are presented. It is shown that the proposed algorithms efficiently reject impulse noise and substantially improve SAR image quality.