A Method for Detecting Sea-Surface Small Targets Based on Features of Time-Varying Polarization Scattering Mechanism

For the detection of small targets in complex sea surface environments, traditional constant false alarm rate (CFAR) detection methods and adaptive detection methods do not perform robustly. In this paper, we propose a tri-feature based detector by utilizing the time-varying differences between the target and sea clutter in polarization scattering mechanism (PSM). First, the three-dimensional data cube of radar echo is constructed by polarization modulation, which includes polarization state, pulse and range information. Second, three features are extracted, composed of polarization stability over time (PST), average decorrelation time (ADT) and enhanced cumulative amplitude (ECA). Finally, integrating the aforementioned three features, a target detector is developed, based on a support vector machine (SVM). Experimental results from the IPIX dataset indicate that, when compared to single-polarization channel detectors, our method outperforms both time-frequency and tri-feature detectors in terms of detection performance. Moreover, in comparison to polarization fusion and polarization decomposition-based multi-polarization channel detectors, our proposed approach also achieves the superior performance.