Fast Detection of Small Targets in High-Resolution Maritime Radars by Feature Normalization and Fusion

Joint exploitation of multiple features is a recognized way to implement the effective detection of small targets on the sea in high-resolution maritime radars at the dwelling mode. The tri-feature-based and feature-compression-based detectors use the time-consuming convex hull learning and decision. In this article, a fast feature-fusion-based detector using seven salient features is proposed, where the convex hull learning and decision are replaced by simple threshold determination and decision. It is found that the seven features of sea clutter can be modeled well by the Burr type XII and t-distributions. From the fitted distribution, each feature is normalized to approximate the standard Gaussian distribution by a nonlinear transformation. In the seven-dimensional normalized feature space, an analytical method is given to calculate the optimal weights of feature fusion by the maximization of the interclass distance of the two-class samples. Owing to the normalization of the seven features, the feature fusion loss is significantly reduced comparing with the direct fusion of seven features. The fast feature-fusion-based detector was evaluated on the open and recognized radar databases and offshore experimental data using an unmanned aerial vehicle with a corner reflector. The results show that the fast detector attains competitive performance with the existing best feature-based detector.