Polarization characteristics of sea clutter and its application in target detection

Abstract

In this paper, three categories of polarization features that represent the divergence between the small target and clutter are discussed and extracted from different views: (1) Since the scattering mechanism of sea clutter is random scattering in low glazing angle, while target has a specific shape, thus reducing the randomness of scattering, the mathematical expectation of the polarization entropy is extracted to reveal randomness of scattering mechanism in the echo; (2) In general, the scattering structure components such as the spherical, dihedral and helicoid scatters are different in the sea clutter and target echoes. Therefore, the power of the three scattering components in each echo is extracted based on Krogager polarization decomposition; (3) Due to the effectiveness of frequency-based and fractal-based features in sea-surface target detection, the average Tsallis entropy is extracted to represent the divergence between sea clutter and small target in full-polarization data. Based on the features in polarization views, localized one-class support vector machine (LOCSVM) is used as the detector because the target feature vector samples are insufficient. The experimental results on IPIX data under different sea conditions show the effectiveness of the proposed method.