Maritime Domain Awareness (MDA): Ship Detection and Classification by Spaceborne Synthetic Aperture Radar

Abstract

In recent years, Maritime Main Awareness (MDA) has attracted much attention. One of the important issues of MDA is ship detection and classification by spaceborne synthetic aperture radar (SAR) for its all-weather and day-and-night imagingcapabilities on a global scale. Here, in this paper, we present examples of ship detection and classification using TerraSAR-X HH/VV dual-polarization data over the Tokyo Bay, Japan and Alboran Sea in Mediterranean Sea. For ship detection,comparison is made with mean image contrast among the different algorithms, including the intensities, inter-polarization coherence, entropy, and HH-VV phase difference. For ship classification, the Ship Monitoring Service (SIMONS) developed by the author (G. M. Martin) is used. Ground-truth data were collected for validation using the simultaneous AIS (Automatic Identification) data and visual observation with a video camera (only AIS data in the Alboran Sea). Although thenumbers of samples were limited, the results of 29 ships showed that the highest image contrast was obtained with coherence (0.92), followed by entropy (0.79), and phase difference (0.73), outperforming the intensity alone (about 0.25). The classification accuracy using 19 ships with AIS data was 80%.