Small Target Detection in Sea Clutter by Time-Frequency Tri-Feature based SVM Detector

Abstract

A SVM detector utilizing three time-frequency features is proposed to discriminate floating small targets in sea clutter environment. The detector consists of three stages. Firstly, transform radar echoes into time-frequency domain. Then, three features, i.e. kurtosis, skewness and concentration, are extracted in the time-frequency domain. Finally, two-class support vector machine (SVM) detector is used to find floating small target under sea clutter background. The result of experiments using real datasets under different circumstances verifies the performance of our proposed method.