Ship Detection in SAR Imagery

 Abstract---As a part of Maritime Domain Awareness, there is a requirement to detect ships in satellite-borne Synthetic Aperture Radar (SAR) images, which provide wide area ocean surveillance. When ship detection is implemented using a Constant False Alarm Rate (CFAR), statistical theory can be employed to ensure that proper parameters are used to find the thresholds for detection; inaccuracy in parameter estimation tends to lead to threshold bias and should be compensated. Otherwise, in spatially varying clutter, the practical performance of automatic ship detectors is likely to be compromised by increases in the false alarm rate and/or by reduction of the probability of detection.