Colour as an attribute for automated detection in maritime environments

This work studies visual detection with camera systems at sea. Electro-Optical sensors are part of a larger picture compilation process and are likely to perform sensor tasks in support of surveillance, classification, identification and anomaly detection. This article investigates the usefulness of colour as an attribute for automated detection in a maritime environment. Analysis of extensive imagery, that was collected from HNLMS Tromp, revealed that the blue colour channel has a higher information density than the other colour channels. For automatic detection of objects at sea, the use of a principal colour channel is investigated. It is concluded that adaptive detection settings with respect to the sensor task and the environment result in an improvement of automatic detection capabilities.