Polarimetric detection of maritime floating small target based on the Complex-valued Entropy Rate Bound Minimization.

Detection of sea-surface small floating targets in maritime high-resolution surveillance radar has been an active area of research in recent years. In this paper, we propose a new detector based on a complex-valued independent component analysis (cICA) algorithm proposed by Geng-Shen Fu et al. called complex entropy rate bound minimization (CERBM), to look for targets in polarimetric radars. It uses received time series at cell under test (CUT) with different polarizations as distinct mixtures. The proposed detector can exploit all information of polarimetric radar for an accurate detection. It does separation on the mixtures using CERBM which results in two output sources, i.e. clutter and target. Finally, the target is detected using estimation of the parameters of K-distribution for outputted sources. Our experiments on the recognized IPIX radar database show that the proposed detector obtains better detection performance in comparison to the newly proposed detectors. The robustness of the detector is also investigated by experiment in either low and high sea state which shows its appropriate results.