Ship feature extraction from maritime radar data

In large harbours and long channels where congested traffic and critical weather are present, VTS (vessel traffic service) systems are needed for traffic management and control. The requirements for these systems become more and more demanding because of the continuous increase of traffic and adverse meteorological conditions; reliable systems which can help the operators to avoid accidents which can jeopardise the security of people and goods have to be used to increase safety and efficiency. These systems employ mainly radar sensors, suitably placed in order to have a complete coverage of the area to be controlled, because a 24-hours operation in all weather conditions must be assured. In this paper a new methodology to process digitised radar images to extract object features (size, heading, etc.) is presented. Considering that ships arriving to, or departing from, harbours are big objects moving in narrow channels or waterways, information such as length and bow position are fundamental to control the ship movements, to evaluate the distance between two vessels, or between the vessel and fixed objects such as buoys or wharves. Results of extensive measurement campaign carried out in Naples, Leghorn and Genoa harbours are reported.