Sea-Surface Floating Small Target Detection Based On Joint Features

This paper proposes a method for sea-surface floating small targets detection. Eight features (relative average amplitude, relative Doppler peak height, relative Doppler spectral amplitude entropy, time-frequency ridge energy, the number of binary connection areas, the largest pass size, relative average power and relative Doppler spectral entropy) are extracted from the radar returns. The detector based on the principle component analysis (PCA) is used to distinguish target bins from sea clutters in the feature space. Experiments using IPIX data sets show that this method achieves better detection performance than several existing detectors.