Fast Dual Trifeature-Based Detection of Small Targets in Sea Clutter by Using Median Normalized Doppler Amplitude Spectra

Abstract

High-resolution sea-surface surveillance radars are designed to find two classes of small targets, low-velocity and floating small targets on the sea surface and low-altitude small targets over the sea surface. Existing feature-based detectors using amplitude and Doppler spectra features of returns can fast detect the first class of small targets. But they often fail to find the second class of small targets due to their small radar cross sections and low signal-to-clutter ratios (SCR). This paper proposes a fast dual tri-feature-based detector to find the two classes of small targets. At the low-velocity tri-feature-based detector, the existing tri-feature-based detector is used to find the first class of small targets. At the high-velocity tri-feature-based detector, a new tri-feature-based detector is used to detect the second class of small targets by utilizing three new features from the median normalized Doppler amplitude spectra (MNDAS). Owing to the ability of the MNDAS to suppress sea clutter, the high-velocity tri-feature-based detector can effectively find the second class of small targets with low SCRs. The detected results in the two tri-feature-based detectors are combined into the binary output of the detector in terms of the ‘OR’ operation. Finally, the proposed detector is verified and compared with the existing detectors by the open and recognized IPIX and CSIR radar databases and one unmanned aerial vehicle radar dataset. Besides the detection performance for the two classes of small targets, a merit of the proposed detector is of low-complexity decision because no time-frequency features are available in detection.