关键词: 数据库管理, 船舶轨迹, 时空立方体, Geodatabase

Abstract:

In order to monitor real-time vessel information to improve navigation safety, China's Maritime Safety Administration (MSA) has built the world's biggest Automatic Identification System (AIS) shore-based network, in which data such as ship position, name, purpose, course and speed are automatic collected 24 hours per day primarily in Chinese coastal waters. As a result, China is approaching the era of big data storage of vessel trajectories, which has brought great challenges to traditional moving objects data management systems. Beyond their basic functions of loading and displaying vessels' position records, an ideal vessel trajectories database should bring user more advance functions of analysis of ship tracking records by supporting spatio-temporal query and prediction of vessel movement. In this paper, we start with the character of vessel movement and abstract the data model of vessel trajectories according to state-of-the-art technology of moving objects databases. Due to the characteristics of vessel trajectories data, such as changing frequently, wide cover range and mass datum, it is argued that current methods of trajectories storage still deserve much more research and improvements, especially for spatio-temporal query and geoprocessing support methods. Considering the role of time perspective, vessel trajectories are managed by three kind of time unit (sampling instant, stepping period and 24 hours) so as to built a three-level organizational framework. By compressing the volume of data and matching original vessel tracking message into spatio-temporal cube unit the retrieval efficiency increases significantly. It was also described how to streamline the acquisition, loading, filtering, display and analysis of raw AIS log files. This method is applied in handling daily mass vessel tracking records which are covering western Taiwan Strait. Experiences show that this model satisfied the requirements of application. The storage is reduced and the performance of spatio-temporal query is improved. Using ArcGIS platform of Geodatabase module, vessel trajectories' initial data model is easy to revised, expanded and integrated with various relational base-map data. Furthermore, it is convenient to apply variable ArcGIS geoprocessing tools to obtain customize demand, such as daily hot spot activities of fishing vessel. By generating these synthesized products our solutions would support the ocean planning community to better understand marine transportation patterns and potential use conflicts between vessels and other activities.

Key words: Geodatabase, spatio-temporal cube, vessel trajectories, database management