Vessel and Oil Spill Detection
Maritime surveillance activities are traditionally carried out by patrol ships or aircrafts. However, the increasing demand for sea and ocean monitoring on a large scale makes this approach unsuitable, due to its cost, its spatial limitations and dependency on weather conditions.
The use of satellite images for ship detection has been widely investigated in the past. The launch of new generation satellites equipped with enhanced SAR systems, able to acquire high quality images at various frequencies, polarisation modes and spatial resolution provides new opportunities for the remote sensing community in the context of maritime surveillance and sea monitoring.
Researchers and students interested in developing ship and oil spill detection algorithms using SAR imagery will greatly benefit from using this research application, which is particularly useful for vessel tracking and oil slick monitoring purposes, when stacks of images are required to be processed in a short time frame.
The ship detection algorithm aims to identify potential vessels on the sea surface, their exact location using geographical coordinates, and characterisation through the retrieval of key parameters, such as size (length and width), route and speed. Ship detection in SAR imagery is possible because ships appear as bright objects over a darker background (water). SAR is also able to detect ship wakes as a series of bright/dark lines, which can supply information about a ship’s direction of movement. Usually, a moving vessel appears distant from its wake. Such displacement due to the inherent nature of SAR image creation, allows the retrieval of the ship velocity.
SAR and optical satellite imagery are also very suitable to detect oil spill at sea. The oil spills in SAR acquisitions typically appear as darker pixels: the density of the oil reduces the roughness of the sea surface, so that the incident microwave signal is mainly scattered away from the antenna. In optical imagery, the spectral signature of the investigated pixels should allow, in principle, to distinguish between sea water and oil spill. It has to be noted that other ocean features reflecting either meteorological or oceanographic conditions can lead to the misinterpretation of the images. Common ‘look-alikes’ in SAR images (i.e. dark structures resembling oil spills) include natural films, low wind surfaces, rain cells, shear zones, internal waves. The oil spill detection algorithm has the objective of detecting and characterise oil spills, discriminating them from ‘look-alikes’. The exact location of the slick, its dimensions and the time of the image acquisition are important to monitor the evolution of the oil spill event.
Combined use of the ship detection and oil spill detection algorithms could be considered to identify the vessel responsible for it, which is key information for national and international maritime safety and surveillance.
Posted on: 2nd June 2016, by : admin