Vessel Motion Detection Using Feature Extraction

Ocean supports human life in many fields, including fishing, resource exploration, transportation, and tourism. On the other hand, numerous challenges need to be addressed, such as the frequent occurrence of maritime accidents, military use, marine pollution, and natural disasters. In order to use the ocean effectively in the future, it is essential to ensure the safety of maritime activities. For this purpose, Maritime Domain Awareness (MDA) is crucial. In MDA, observations and the detection of objects on the ocean are carried out using satellites, aircraft, and other platforms. The objective is to monitor environmental changes and maritime activities by utilizing the obtained data. Especially in monitoring maritime activities, the detection of vessels is essential. Furthermore, in order to enable early response to suspicious vessels, it is necessary to detect not only the position and the type of the vessel but also its motion. However, much of the research focuses on the identification of the position and the type of vessel, with limited research on the vessel's motion detection. Additionally, much of the research on vessel motion detection involves using object tracking, which requires multiple continuous images of the vessel. Therefore, the purpose of this paper is to propose a method to detect the vessel motion from a single image.