Ship heading and velocity analysis by wake detection in SAR images

• Published in: Acta astronautica Vol. 128; pp. 72 - 82
• Main Authors: Graziano, Maria Daniela, D’Errico, Marco, Rufino, Giancarlo
• Format: Journal Article
• Language: English
• Published: Elmsford Elsevier Ltd 01.11.2016; Elsevier BV
• Subjects: Algorithms; Angular velocity; Azimuth; Image detection; Incidence angle; Inspection; Object detection; Radar imagery; Radar imaging; Radon Transform; Remote Sensing; Satellite tracking; Ship-route estimation; Ship-wake detection; Ships; Spaceborne radar; Synthetic Aperture Radar; Threshold logic; Turbulence; Turbulent flow; Velocity; Wakes; Italy; MED, Italy, Campania, Napoli, Napoli Gulf; Ship-route estimation; Synthetic Aperture Radar; Spaceborne radar; Ship-wake detection; Object detection; Remote Sensing; Radon Transform
• ISSN: 0094-5765; 1879-2030
• DOI: 10.1016/j.actaastro.2016.07.001

With the aim of ship-route estimation, a wake detection method is developed and applied to COSMO/SkyMed and TerraSAR-X Stripmap SAR images over the Gulf of Naples, Italy. In order to mitigate the intrinsic limitations of the threshold logic, the algorithm identifies the wake features according to the hydrodynamic theory. A post-detection validation phase is performed to classify the features as real wake structures by means of merit indexes defined in the intensity domain. After wake reconstruction, ship heading is evaluated on the basis of turbulent wake direction and ship velocity is estimated by both techniques of azimuth shift and Kelvin pattern wavelength. The method is tested over 34 ship wakes identified by visual inspection in both HH and VV images at different incidence angles. For all wakes, no missed detections are reported and at least the turbulent and one narrow-V wakes are correctly identified, with ship heading successfully estimated. Also, the azimuth shift method is applied to estimate velocity for the 10 ships having route with sufficient angular separation from the satellite ground track. In one case ship velocity is successfully estimated with both methods, showing agreement within 14%. •A wake detection method is developed and applied to X-Band SAR images.•The algorithm foresees an innovative Radon-based detection of the wake features.•Wake features are validated as real wake signatures by merit indexes.•The method is tested over 34 ship wakes in both HH and VV images.•The azimuth shift method is applied to estimate ship velocity.