基于数据挖掘的船舶行为研究  

Research on Ship Behaviors Based on Data Mining

作　　者：朱飞祥[1] 张英俊[1] 高宗江[1] 

机构地区：[1]大连海事大学航海学院,辽宁大连116026

出　　处：《中国航海》2012年第2期50-54,共5页Navigation of China

基　　金：国家自然科学基金项目(61073134);中央高校基本科研业务费专项资金资助项目(2009QN009);(2012QN003)

摘　　要：如何从我国沿海港口海量的船舶自动识别数据中找出海上交通知识,是我国交通海事部门和港口管理部门亟需解决的重要问题。针对海量数据处理的瓶颈及海上交通数据的特点,运用地理网格技术划分港口水域,从而降低船舶航迹数据复杂度,建立船舶行为模型,对每个地理单元格上的船舶行为进行统计分析,进一步运用数据挖掘技术揭示整个海域船舶行为规律。所挖掘的知识可以运用到船舶航行位置预测、船舶异常行为检测及海上交通流模拟等研究领域。该研究方法开拓了海上交通安全的研究思路,为海事部门和港口管理部门的通航环境管理等提供理论依据。How to find out hidden marine traffic knowledge from huge amount of automatic identification system （AIS） data of China＇s harbors is a problem to be solved by maritime safety administration and harbor authorities. In light of the volume of data and the feature of traffic data, geographic grid technology is introduced to reduce the complication of ship trajectories. The model of ship behaviors is set up and the statistics and analysis of the ships＇ behaviors are carried out on a geographic grid basis. The ship behavior knowledge in whole harbor water area is re- vealed by further processing with the data mining technology. The mined knowledge can be used in ship position prediction, ship abnormal behavior detection and traffic flow simulation. This study demonstrates a new approach to investigate traffic safety at sea and forms a theoretical basis of navigational environment management for maritime and harbor administrations.

关 键 词：水路运输 海上交通 船舶行为 数据挖掘 船舶自动识别系统 

分 类 号：U698[交通运输工程—港口、海岸及近海工程]