基于时空密度的船载AIS数据聚类分析方法研究  

Clustering Analysis Method of Shipborne AIS Data Based on Spatio-Temporal Density

作　　者：李永攀 刘正江[1] 郑中义[1] LI Yongpan;LIU Zhengjiang;ZHENG Zhongyi(Navigation College,Dalian Maritime University,Dalian 116026,Liaoning,P.R.China;Vessel Traffic Management Center,Zhoushan Maritime Safety Administration,Zhoushan 316000,Zhejiang,P.R.China)

机构地区：[1]大连海事大学航海学院,辽宁大连116026 [2]舟山海事局船舶交通管理中心,浙江舟山316000

出　　处：《重庆交通大学学报（自然科学版）》2018年第10期117-122,共6页Journal of Chongqing Jiaotong University(Natural Science)

摘　　要：船载AIS数据包含位置、时间和其他属性,属于典型的时空数据,对其开展时空聚类分析有助于挖掘海上交通特征。结合AIS数据的具体特征,提出时间切片化方法,有效约简AIS数据并处理报告间隔不等的问题。在DBSCAN算法基础上综合考虑时间和空间要素,提出船载AIS数据时空聚类算法,并对实际数据开展分析。该方法能更好顾及船舶交通流的时空耦合特征,识别隐含的时空模式,为主管机关开展船舶交通管理、优化通航秩序、保障航行安全等提供一种新途径。Shipborne AIS data contains location,time and other attributes,which belongs to typical spatio-temporal data.And the spatio-temporal clustering analysis on shipborne AIS data contributes to the data mining of maritime traffic characteristics.Considering the specific characteristics of AIS data,a time slicing method was proposed to effectively simplify the AIS data and deal with the problem of different report intervals.Based on DBSCAN algorithm,combining time and space attributes,a spatio-temporal clustering algorithm of shipborne AIS data was put forward,and the actual AIS data was taken for a case study.The proposed method can better take account of the spatial-temporal coupling characteristics of ship traffic flow and identify the hidden spatio-temporal pattern,which provides a new way for the competent authorities to carry out ship traffic management,optimize navigation order and ensure navigation safety.

关 键 词：交通运输工程 时空密度 聚类分析 DBSCAN算法 船载AIS 

分 类 号：U675.79[交通运输工程—船舶及航道工程]