中国安全科学学报 ›› 2019, Vol. 29 ›› Issue (5): 111-116.doi: 10.16265/j.cnki.issn1003-3033.2019.05.019

• 安全工程技术科学 • 上一篇    下一篇

基于AIS的数据时空分析及船舶会遇态势提取方法

马杰1,2,3 副教授, 刘琪1, 张春玮1, 刘克中1,2,3 教授, 张煜3,4 教授   

  1. 1 武汉理工大学 航运学院,湖北 武汉 430063;
    2 内河航运技术湖北省重点实验室,湖北 武汉 430063;
    3 国家水运安全工程技术研究中心,湖北 武汉 430063;
    4 武汉理工大学 物流学院,湖北 武汉 430063
  • 收稿日期:2019-02-02 修回日期:2019-04-03 发布日期:2020-11-02
  • 作者简介:马杰(1978—),男,湖北武汉人,博士,副教授,主要从事大数据挖掘、智能航海方面的研究。E-mail:majie@whut.edu.cn。
  • 基金资助:
    国家自然科学基金资助(51679182,71874132);智能船舶1.0-船舶辅助自动驾驶系统开发项目(工信部联装函〔2016〕544号)。

A method for extracting ship encounter situation based on spatio-temporal analysis of AIS data

MA Jie1,2,3, LIU Qi1, ZHANG Chunwei1, LIU Kezhong1,2,3, ZHANG Yu3,4   

  1. 1 School of Navigation,Wuhan University of Technology,Wuhan Hubei 430063,China;
    2 Hubei Inland Shipping Technology Key Laboratory,Wuhan Hubei 430063,China;
    3 National Engineering Research Center for Water Transportation Safety, Wuhan Hubei 430063, China;
    4 School of Logistics Engineering, Wuhan University of Technology,Wuhan Hubei 430063,China
  • Received:2019-02-02 Revised:2019-04-03 Published:2020-11-02

摘要: 为精准提取船舶会遇态势,提升水上交通安全监管能力,对长江口南槽水域自动识别系统(AIS)数据做时空分析,提出对船舶会遇态势模式分类的自动提取方法。首先,利用会遇态势过程的时空约束关系提取会遇船舶配对轨迹信息;然后,借助数据插值方法对会遇轨迹做时空同步处理和数据补全,实现会遇场景重建;最后,分析船舶会遇的时空演化特征,提取特定时间窗口内的相对距离和航向差特征,形成会遇特征序列,利用支持向量机(SVM)对会遇特征序列分类辨识建模,实现会遇态势的自动提取。结果表明:设置时空约束条件可以准确提取船舶配对轨迹信息;对会遇过程作时空分析,实现了会遇场景的重建;借助SVM设计的会遇态势提取算法的准确率达90%以上,与传统方法相比降低了误判率。

关键词: 船舶会遇态势, 自动识别系统(AIS), 时空分析, 支持向量机(SVM), 自动提取

Abstract: In order to accurately extract ship's encountering situation, spatial and temporal analysis of AIS data from the south channel of the Yangtze river estuary was carried out, and an automatic extraction method for the encountering situation of ships based on pattern classification was proposed. Firstly, the spatio-temporal constraint conditions in the encounter process were used to extract the ship matching information. Then encounter trajectory was synchronized by data interpolation, and encounter scene was reconstructed. Finally, spatio-temporal evolution characteristics of ship encountering were analyzed, and relative distance and course difference features in a specific time window were extracted to form encounter feature sequence. SVM algorithm was used to classify and identify the encounter feature sequence to realize the automatic extraction of encounter situation. The results show that setting spatio-temporal constraints can accurately extract ship pairing trajectory information, that the spatio-temporal analysis of encounter process can reconstruct encounter scene, and that the accuracy rate of the encounter situation extraction algorithm designed by SVM is over 90%, which reduces the misjudgment rate compared with traditional methods.

Key words: ship encounter situation, automatic identification system(AIS), spatiotemporal analysis, support vector machine(SVM), automatic extraction

中图分类号: