DECLIMS Detection and Classification of Maritime Traffic from Space

Finding ships from space Space-based imaging enables the surveillance of the planet’s oceans for maritime traffic, increasingly in nearreal-time. The challenges are in analysing the satellite images: to find ships and their positions; to determine what kind of ships they are; to estimation their speed and course; and to deduce their activity. We have two types of sensors at our disposal: Synthetic Aperture Radar (SAR) and optical. The SAR can image very large areas (up to 400 km swath width) and works independently of cloud cover or daylight. However, it cannot provide much detail about the targets it finds: classification capability is limited. For this we need optical imagers. These can show details as small as 1 meter or less, making classification much easier. But their operation is limited to clear, daytime conditions, and their images cover much smaller areas. The two satellites that have been mostly used for ship detection are the European ENVISAT and the Canadian RADARSAT, both SAR sensors. Optical satellites have been used to a lesser extent: the main ones include SPOT (France), IKONOS, QUICKBIRD (US) and EROS (Israel). Between them they offer the possibility to survey large areas or to focus on small areas for details. These satellites orbit the Earth at some 700 km altitude. Due to this orbiting, it is not possible to continuously monitor a certain location (as opposed to the capacity of a geostationary satellite). From space, we therefore obtain snapshots of the situation, and the revisit frequency to a certain area is at present normally once every few days (more frequent if more satellites are used, and for areas closer to the poles).