Statistical Characterization of High Resolution Sar Images under Different Sea State Conditions

Accurate ship detection from radar imagery needs a proper statistical characterization of both target and sea scattering. Even if a statistical model for the target class is not available, effective discrimination between “sea” and “nonsea” classes can be achieved making use of a good model for the sea scattering distribution (i.e. anomaly detection). Hence, the retrieval of the statistical characterization of sea is a very important task for target detection over the sea surface. In this paper we analyze the statistical characterization of sea surface for high resolution SAR imagery. This topic has been poorly investigated by the scientific community, especially for data acquired by military or dual use (military and civil) radar system such as Cosmo/Skymed (C/S) satellite constellation and TELAER (TELerilevamento AERonautico) airborne system.