王震 李伟峰 高邈

摘要：為提高船舶进出交通流密集区域的安全性、解决数据挖掘不充分的问题，基于AIS数据，将多种算法相结合，提出一种多元化的船舶交通流框架提取方法。利用Douglas-Peucker压缩算法和航迹交会算法分别提取交通流中的船舶转向点和航迹交会点。利用密度聚类算法对包括船位点在内的3种特征点进行数据挖掘，提取出更有代表性的特征点。将3种特征点进行加权融合，得到新的多元特征点，以点的大小表示其重要程度，最终生成某水域的船舶交通流框架。实验结果表明，通过以上方法能够获得老铁山水道附近水域船舶交通流框架。该框架融合了多种航迹特征点，能够显示附近水域的重要航迹分布，充分体现船舶交通流的总体态势和密集区域;该框架从统计学角度凝结了该水域船舶行驶的习惯航线，这些航线具有较好的适航度，既可用于航路规划，还能为海事部门选取推荐航道提供参考。

关键词：

数据挖掘; 船舶交通流; 特征点; 船舶自动识别系统（AIS）

中图分类号：  U692.37

文献标志码：  A

收稿日期： 2020-09-07

修回日期： 2020-12-16

基金项目：

中央高校基本科研业务费专项资金（3132020134，3132020139）

作者简介：

王震（1996—），男，山东聊城人，硕士研究生，研究方向为AIS大数据挖掘，（E-mail）1506216436@qq.com;

李伟峰（1983—），男，山东菏泽人，副教授，硕士，研究方向为船舶智能避碰，（E-mail）sddmlwf@163.com

Framework extraction of ship traffic flow with

multi-algorithm combination

WANG Zhen， LI Weifeng， GAO Miao

（

Navigation College， Dalian Maritime University， Dalian 116026， Liaoning， China）

Abstract：

In order to improve the safety of ships entering and leaving traffic-intensive waters and to solve the problem of insufficient data mining， a diversified method for extracting the framework of ship traffic flow is proposed based on AIS data and with the combination of multiple algorithms. The Douglas-Peucker compression algorithm and the trajectory crossing algorithm are used to extract the ship turning points and the trajectory crossing points in the traffic flow. The density clustering algorithm is used to conduct data mining on the three types of characteristic points including the ship position points， so as to extract more representative characteristic points. The three types of characteristic points are weighted and fused to obtain new multivariate characteristic points， and the framework of ship traffic flow in a certain waters is generated， in which the size of a point represents the importance. The experimental results show that the framework of ship traffic flow in the local waters of Laotieshan channel can be obtained through the above method. The framework integrates a variety of trajectory characteristic points， which can display the distribution of important trajectories nearby and fully reflects the overall situation and dense areas of ship traffic flow. It also condenses the customary routes of ships in the waters statistically， and the customary routes have good seaworthiness and can be used for route planning and reference for maritime departments to select recommended channels.

Key words：

data mining; ship traffic flow; characteristic point; automatic identification system （AIS）

0 引 言

随着船舶自动识别系统（automatic identification system，AIS）的广泛使用，海事系统及船公司接收了大量包括船舶航迹及海上交通环境等多种信息在内的AIS数据。为获取AIS数据中蕴藏的船舶交通流和航行环境的特征及规律，运用大数据算法对其进行数据挖掘已成为一个重要研究方向。……