基于统计学理论的船舶轨迹异常识别  

A Study on the Identification of Abnormal Ship Trajectory Based on Statistic Theories

作　　者：甄荣[1] 邵哲平[1] 潘家财[1] 赵强[1] 

机构地区：[1]集美大学航海学院,福建厦门361021

出　　处：《集美大学学报（自然科学版）》2015年第3期193-197,共5页Journal of Jimei University：Natural Science

基　　金：福建省自然科学基金资助项目(2012G0030)

摘　　要：为了准确识别异常航行轨迹的船舶,以船舶AIS信息为数据源,利用统计学中曲线拟合的最小二乘法对训练集船舶轨迹点进行拟合,得到船舶典型航行轨迹的数学表达模型,以此作为标准,通过计算监控船舶轨迹点与典型轨迹间的距离是否大于典型轨迹95%置信区间的范围,从而对轨迹异常的船舶进行识别.实验结果表明,该方法可以有效地识别轨迹异常船舶.将该方法运用到监控系统中可以提高海上交通监控效率.In order to identify the abnormal trajectory of a ship in the maritime traffic monitoring system automatically, sourced from shipborne AIS data, the curve fitting of statistic theories was used to fit the traj- ectory points of a training vessel, so as to obtain the mathematical model for the typical route of a ship. Based on this, ships with abnormal trajectories are identified through calculating whether the distance between valida- tion data and typical route is wider than the 95% confidence interval of the typical route. The results show that abnormal ship trajectories can be identified efficiently this way. If this is applied to the vessel traffic sur- veillance system, the efficiency such surveillance wiU be increased significantly.

关 键 词：船舶 航行轨迹 统计学 AIS信息 监控 

分 类 号：U675.7[交通运输工程—船舶及航道工程] TP391[交通运输工程—船舶与海洋工程]