A New Approach for Ship Detection in SAR Imagery Based on Convolution between Different Polarization Channels

Abstract

A new approach is proposed for ship detection with SAR (Synthetics Aperture Radar) images based on the convolution between different polarimetric channels. The method takes advantages of different statistical behaviors among ships and surrounding ocean, interpreting the information through the convolution in order to provide a more reliable detection. The simulations show that the convolution is sensitive to the size and shape of ships not to sea state. The method can be used under high sea state and have no relation to the movement of ships. Moreover, the method is general which is not only suitable for dual polarization data but also for full polarization data. The detection performance over both simulated and real images confirms the robustness of the method.