Detection of Small Floating Targets on the Sea Surface Based on Multi-Features and Principal Component Analysis

This letter proposes a detection method of small floating targets based on multi-features and principal component analysis (PCA) for marine surveillance radar. This method consists of three stages. In the first stage, six features extracted from radar returns are combined into a feature vector. Numerous feature vectors of sea clutter make up a feature matrix. In the second stage, the feature matrix is decomposed into low-rank and sparse components to decrease the influence of sea spikes. In the third stage, a PCA-based anomaly detector with an adjustable false alarm rate is constructed to distinguish the target cells from the clutter-only cells in a feature space because prior information of the target is usually unknown. Experiments using the measured database of an ${X}$ -band radar show that the proposed method attains a state-of-the-art detection rate.