基于AIS数据的船舶会遇挖掘与分析

Mining and Analysis of Ship Encounters Based on AIS Big Data

为实现在AIS海量数据中快速高效地识别会遇船舶信息,对会遇船舶航行过程实施监控和分析,识别研究水域内存在的海上交通安全风险点。以台湾海峡部分区域内的AIS数据作为研究对象,运用改进的DBSCAN聚类算法,结合船舶会遇特征参数计算,挖掘两船会遇与多船会遇信息,根据船舶方位和航向差划分船舶交叉、对遇和追越三种会遇局面,并通过定量化方式多维度描述不同类型的船舶会遇过程。结果表明:分析轨迹数据的时空特性可以再现船舶会遇场景,能更清晰全面地展现船舶的会遇过程。

In order to realize the rapid and efficient identification of encountering ships in the massive data of AIS,the research monitors the navigation process of encountering ships by?using an improved DBSCAN clustering algorithm,which?is?well used in rapid extraction of maritime traffic safety risks in the research water area.Through the combination of ship encounters feature parameter calculations and using AIS data within the Taiwan Strait area as the research object,dual ship encounters and multi-ship encounters are mined effectively.Then the encounter situations are divided into three types,ie,cross encounters,head-on encounters and overtaking encounters,based on the ships’bearing and course differences.Finally,the encounter processes of different types of ships are described in multiple dimensions,the encounter scenes of the ship are reproduced through a quantitative?method and time-space characteristics?analysis.The?results?show the analysis of time-space chatacteristics can reproduce ships encounter situations and illstrustrate the encounter process more clearly and comprehensively.