首页 | 本学科首页   官方微博 | 高级检索  
     检索      

船舶进离港行为捕获算法设计与实现
引用本文:王永明,饶全成,姚萍.船舶进离港行为捕获算法设计与实现[J].上海海事大学学报,2018,39(1):6-12.
作者姓名:王永明  饶全成  姚萍
作者单位:大连海事大学航海学院,中国科学院计算技术研究所,中国科学院计算技术研究所
基金项目:国家自然科学基金(51309044)
摘    要:针对在设计船舶进离港行为捕获算法时存在港口范围难定义、AIS数据的经纬度误差造成船舶频繁进出港区边界等问题,将全球AIS数据与港口数据相结合,设计船舶进离港行为捕获算法。根据船舶速度、位置、运动特性等信息,实现对船舶进离港、停泊等行为的判断。利用纬度分桶法、网格法和Kd树法实现基于船舶位置的港口匹配。实际数据测试表明,网格法的港口匹配时效性最好,每秒可处理约16万条AIS数据。随机选取30艘船对算法进行验证,结果表明船舶轨迹与捕获到的船舶进离港事件相符。

关 键 词:船舶行为    自动识别系统(AIS)    港口匹配    进离港
收稿时间:2017/5/22 0:00:00
修稿时间:2017/9/18 0:00:00

Design and implementation of capturing algorithm for ship entering and leaving port behavior
Institution:China Transport Telecommunications & Information Center,Navigation & Informatization Institute.,Institute of Computing Technology, Chinese Academy of Science and Institute of Computing Technology, Chinese Academy of Science
Abstract:In the design process of capturing algorithm for ship entering and leaving port behavior, it is difficult to define port scope, and ships go in and out of port boundary frequently due to latitude and longitude errors of AIS data. To solve these problems, global AIS data and port data are combined to design an algorithm for capturing ship entering and leaving port behavior. The velocity, position and motion characteristics of ships are used to identify ship behaviors such as entering and leaving port, mooring. The latitude point barrel method, the grid method and the Kd tree method are used to realize port match based on ship position. The real data test shows that the grid method is optimal in time and effect for port match, and it can handle about 160 000 messages in 1 s. Thirty ships are selected randomly to verify the algorithm, and the result shows that the ship tracks accord with the events of entering and leaving port of captured ships.
Keywords:ship behavior  automatic identification system (AIS)  port match  entering and leaving port
本文献已被 CNKI 等数据库收录!
点击此处可从《上海海事大学学报》浏览原始摘要信息
点击此处可从《上海海事大学学报》下载免费的PDF全文
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号