Spatiotemporal Data Mining And Visualization Of Ship Trajectories Based On AIS Data

As an important node of the "21st-century maritime silk road",the Pearl River Estuary is a natural barrier and strategic channel for China’s national security.AIS(Automatic Identification System)data contains a large amount of ship information and navigation time and space information,which is important data for regional water transportation assessment.Based on the AIS data,the water traffic characteristics of ships in the Pearl River Estuary were mined,which provided theoretical support for optimizing the water transport network,improving the transport efficiency,fighting against water crimes and preventing traffic accidents.The current research mainly focuses on the spatial characteristics of water traffic,but mostly ignores the temporal characteristics of water traffic,and rarely excavates the spatial and temporal characteristics of ship trajectory,trajectory anomaly detection and port spatial information in the Pearl River Estuary.For this reason,based on AIS data,this paper analyzes the space-time characteristics of water traffic in the Pearl River Estuary,identifies key areas of water,detects outlier locus points,and analyzes the hierarchical structure of the port by using computer technologies and methods such as space-time division,linear density analysis,matrix decomposition,GIS technology,cluster analysis,and complex network.The main contents and conclusions are as follows:(1)Spatial and temporal feature mining and visualization of water traffic.The spatial and temporal characteristics of water traffic are analyzed based on the methods of space-time division,linear density analysis,statistical analysis and matrix decomposition.The study found that there are 1 primary waterway and 3 secondary waterways.According to the number of routes,each area is divided into busy area,busier area and less busy area.According to the curve of route volume variation,the regions are divided into bimodal,unimodal and stationary.According to the change of the route volume with time,each area is divided into all-day busy,evening busy and morning busy.(2)Identification of key areas and detection and visualization of trajectory anomalies.Algorithms such as hierarchical clustering and DBSCAN clustering are used to mine key areas and detect abnormal trajectories.The study found that large ports and sea-going/sea mouth are important areas;The spatial distribution of berths of different ship types is mainly affected by the natural environment,geographic location and economic development level of the area;The clustering algorithm can effectively detect outlier trajectory points and find out ships with abnormal spatial position and ship speed in time.(3)Port spatial information mining and visualization.The study found that the ports with the large port degree,point degree and centrality are Hong Kong port and Guangzhou port,while the smaller ports are Zhuhai port,Humen port and Foshan port;Pearl River Estuary ports can be divided into five levels.Hong Kong port,Guangzhou port and Shenzhen Shekou port are at the first level,which is the hub port of Pearl River Estuary.And the centrality difference of different level port is obvious.