Detection of Abnormal Ship Trajectory Points Based on Statistical Learning

Based on the multi-source heterogeneous Marine big data such as ship files, ship tracks and electronic charts, it is helpful to clarify the navigation rules of specific ship types and scales, facilitate the regulatory authorities to accurately perceive the navigation situation of ships at sea, and ensure the navigation safety of coastal countries. Through the method of statistical learning, the paper excavates the rules of the trajectory point position, speed and heading of specific ships in a specific navigation area, sets the position deviation threshold, speed deviation threshold and heading deviation threshold, constructs the detection model of ship abnormal behavior, and realizes the rapid identification of abnormal behavior of ships. The experimental results show that the statistical learning method can better identify the abnormal trajectory points according to the three set thresholds and realize them.