| 一种顾及时间特征的船舶轨迹DBSCAN聚类算法 |
| |
| 引用本文： | 郭乃琨,陈明剑,陈锐.一种顾及时间特征的船舶轨迹DBSCAN聚类算法[J].测绘工程,2021,30(3):51-58. |
| |
| 作者姓名： | 郭乃琨 陈明剑 陈锐 |
| |
| 作者单位： | 信息工程大学 地理空间信息学院,河南 郑州 450001;92493部队,辽宁 葫芦岛 125001;信息工程大学 地理空间信息学院,河南 郑州 450001 |
| |
| 基金项目： | 国家自然科学基金资助项目(41604011)。 |
| |
| 摘 要： | 随着世界海洋经济的快速增长和各国海洋贸易的持续发展,船舶AIS系统被世界各国广泛采用,由此产生海量的船舶轨迹数据。如何对这些多维、动态的数据进行挖掘和利用,成为当前时空数据挖掘领域的研究热点之一。文中在经典DBSCAN空间聚类算法的基础上,对船舶轨迹数据进行清洗、压缩等预处理,并将其划分为特征点相连的子轨迹段,然后引入时间距离度量方法,实现对船舶轨迹的时空聚类。最后基于东海某海域(113°45′37″E~130°23′43″E,17°47′29″N~38°52′59″N)近一个月的船舶轨迹数据进行实验,结果表明相比经典DBSCAN算法,文中算法能够在兼顾时间信息的基础上,对船舶轨迹数据进行有效的时空聚类,为后续研究预测船舶的行为模式奠定基础。
|
| 关 键 词： | 时间特征 船舶轨迹 AIS DBSCAN 时空聚类 |
A DBSCAN clustering algorithm of ship trajectory considering time characteristics |
| |
| GUO Naikun,CHEN Mingjian,CHEN Rui.A DBSCAN clustering algorithm of ship trajectory considering time characteristics[J].Engineering of Surveying and Mapping,2021,30(3):51-58. |
| |
| Authors: | GUO Naikun CHEN Mingjian CHEN Rui |
| |
| Institution: | (College of Geospatial Information, Information Engineering University, Zhengzhou 450001, China;Troops 92493, Huludao 125001, China) |
| |
| Abstract: | With the rapid growth of the world’s marine economy and the continuous development of ocean trade in various countries,the ship AIS system has been widely adopted by countries all over the world,resulting in massive amounts of ship trajectory data.How to mine and utilize these multi-dimensional and dynamic data has become one of the current research hotspots in the field of spatiotemporal data mining.Based on the classic DBSCAN spatial clustering algorithm,this paper preprocesses the ship trajectory data such as cleaning and compression,divides them into sub-trajectory segments connected by feature points,and then introduces a time distance measurement method to realize the ship trajectory spatio-temporal clustering.Finally,experiments are conducted based on ship trajectory data in a certain area of the East China Sea in the past month,whose longitude ranges from 113°45′37″E to 130°23′43″E,and whose latitude ranges from 17°47′29″N to 38°52′59″N.The result shows that compared with the classic DBSCAN algorithm,this algorithm can effectively cluster the ship trajectory data on the basis of time information,and predict the ship’s performance for subsequent research to lay the foundation for behavioral patterns. |
| |
| Keywords: | temporal characteristics ship trajectory AIS DBSCAN spatio-temporal clustering |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
