Extracting ship stopping information from AIS data

Ship stopping information is important for ship voyage division, ship traffic flow analysis, and maritime trade network analysis. A method of ship stopping information extraction that integrates trajectory features and geographic scenario semantics is proposed based on automatic identification system (AIS) data. First of all, under the constraints of port geographic knowledge, the ship trajectory features are excavated to realize the recognition of ship stopping points in the port area with the recognition accuracy rate of 0.94, the recall rate of 0.91, and the F-1 value of 0.92. Secondly, 15 ship stopping mode classification features are selected from both trajectory features and geographic semantics, and a random-forest-based ship stopping classification model is constructed. Two stopping modes, berth stopping and anchorage stopping, are classified with the overall classification accuracy rate of 0.93 and the Kappa coefficient of 0.87. Finally, the port points are generated based on the ship berth stops, and 1,050 port points in the South China Sea Silk Road region are extracted with an accuracy rate of 98.41%. The experimental results show that the proposed method can effectively extract ship stopping information and provide knowledge support for ship behavior understanding and ship traffic analysis.

Automated summary

This study proposes a method for extracting ship stopping information based on ship trajectory features and geographic semantics. By excavating trajectory features, the stopping points in the port area are recognized, and a classification model for ship stopping is constructed. Experimental results show that this method can effectively extract ship stopping information and provide support for ship behavior understanding and ship traffic analysis.