A framework of moving behavior modeling in the maritime surveillance

Due to requirements of homeland security, maritime traffic surveillance becomes an important research area. An automated solution to analyze the collected data and summarize them into a model is helpful for objects' movement behavior analysis in maritime domain. However, it is a challenging problem to develop a model that can precisely capture the objects' movement behavior due to objects would seldom exactly repeat the same moving path in the maritime space, a free moving space. In this paper, based on multiple vessels' trajectories from Automatic Identification System (AIS), we aim to mine trajectory patterns and profile them into a model, called TMP model (Trajectory Pattern Mining and Profiling model). We conduct experiments on real data and experimental results demonstrate that TMP model can effectively and precisely reflect the movement behaviors in the maritime surveillance.