Findings of the DECLIMS project – Detection and classification of marine traffic from space
Via Fermi 1,
21014 Ispra (VA),
The DECLIMS FP5 concerted action brings together 24 parties active in space-based ship detection; from industry, academia and organizations; civilian and defence backgrounds. The project is characterised by exchange of experiences, and comparing performances of algorithms and systems by running benchmark tests. This presentation intends to report on the main findings (so far) of the project which has been running for over two years. These findings comprise the following. Main users of space-based ship detection are fisheries control and defence authorities; civilian security use seems upcoming. The combination of satellite SAR derived vessel positions with those from AIS (vessel transponder) is gaining much attention recently due to the broad uptake of the latter. Automatic ship detection software implemented for fast processing has quite good properties in terms of detection and false alarm rates, but still not as good as a human operator who is better at dealing with complex clutter situations. In combination with a good communication infrastructure, detected ship positions can be at the users within 30 minutes after acquisition. Many systems that call themselves operational have only recently installed the capacity for giving size estimates with their detections. Still, in many cases this estimate is rendered worthless on account of the azimuth smearing that the target, moving on the swells, is subject to. No operational system seems to be able to venture a guess as to the vessel type. The optimum way to deal with ENVISAT ASAR AP mode images is still under debate, and has not reached most operational systems yet. ENVISAT ASAR WSM receives some interest to detect larger (cargo) shipping, but its resolution makes it of very limited use for detection of (smaller) fishing vessels. With regard to future systems, maritime surveillance applications in most cases will favour swath width over polarimetric information. Full polarimetry is likely to be at best confined to coastal or port studies. The limited classification potential means that spaceborne SAR should primarily be used as a cueing sensor, implying the need for very fast turnaround and tasking capability.