Ship Detection in Ice-Infested Waters Based on Dual-Polarization SAR Imagery

This letter discusses the potential of automatic ship detection in ice-infested waters based on satellite synthetic aperture radar (SAR) imagery. The popular
K
-distribution is used to model the backscatter statistics of sea ice clutter. The goodness of fit of this model is assessed with the Kolmogorov-Smirnov and Anderson-Darling test statistics for both VV and VH polarizations. We also test the impact of introducing the Method of Log Cumulant (MoLC) estimator for the shape parameter of the
K
-distribution. Finally, a constant false-alarm rate ship detection algorithm, applying the
K
-distribution with the MoLC estimator, is evaluated on dual-polarization RADARSAT-2 SAR data. Our results demonstrate that this is a viable approach to ship detection in ice-infested waters.