SAR Clutter Modelling in Complex Images for Ship CFAR Detection

A ship detector exploiting complex Synthetic Aperture Radar (SAR) data is designed based on a statistical characterization of sea clutter. SAR satellite sensors are powerful tools for Maritime Domain Awareness. Recent studies show that complex data in single-channel SAR imagery could be underrated. SAR images, acquired by PAZ, in maritime scenarios are used for determining the clutter model that better fits the variations of real and imaginary parts. The goodness-of-fit test confirms a Gaussian model when only one dimension is considered. Then a detector based on two Constant False Alarm Rate (CFAR) techniques, applied in rows and columns respectively, is designed. Order statistics-CFAR, which presents better performance than cell averaging-CFAR, shows the capability of the proposed solution to maritime traffic monitoring without the generation of detected products.