On the novel use of model-based decomposition in SAR polarimetry for target detection on the sea

Abstract

In this study, we show the novel applications of model-based scattering power decomposition analyses in synthetic aperture radar (SAR) polarimetry for man-made target detection on the sea surface. Model-based decomposition technique is primarily used for land-cover classification mainly because microwave scattering from land is composed of various scattering mechanisms. On the other hand, this technique has not been widely used for oceanic applications since the scattering from sea is mostly surface scattering, and obtaining a classification of different types of scattering is not as important as that on land. However, if an object is present on the sea surface, which gives rise to different scattering characteristics from the sea surface, the decomposition approach may be a useful technique for the detection and classification of the object. We suggest two approaches for target detection on the sea surface. One is to use the model-based decomposition as a polarimetric band-stop filter to block the dominant scattering from the background sea surface, and the other is to focus on the optimized double-bounce scattering component associated with the target using a rotation scheme of the polarimetric matrix. The advantage of these methods is the simplicity of the concept and the algorithm, in which no target-centric analysis is necessary. Experimental results show that model-based decomposition analyses can work as powerful target detectors, effectively separating scattering by targets from dominant background scattering from the sea surface.