舶舶AIS轨迹异常的自动检测与修复算法  

Automatic Detection and Restoration Algorithm for Trajectory Anomalies of Ship AIS

作　　者：吴建华[1,2] 吴琛[1,2] 刘文[1,2] 郭俊纬 

机构地区：[1]武汉理工大学航运学院,武汉430063 [2]内河航运技术湖北省重点实验室,武汉430063

出　　处：《中国航海》2017年第1期8-12,101,共6页Navigation of China

基　　金：国家自然科学基金(51279151)

摘　　要：船舶自动识别系统(Automatic Identification System,AIS)轨迹是船舶位置和时间的记录序列,对船舶航行状态分析、船舶交通流预测及船舶海事事故分析等具有一定的意义,而受一些因素影响,采集到的AIS数据会产生错误,导致AIS轨迹出现异常。对此,在对大量AIS数据进行深入解析的基础上,归纳出AIS轨迹异常的几种类型,针对各类型的特征对船舶轨迹异常进行自动检测,并采用3次样条插值模型实现修复算法。研究成果可为基于船舶航行轨迹数据的相关研究提供快速、准确的技术支持。The AIS（ Automatic Identification System） trajectory of ship navigation, the recorded sequence of the location and the time of a ship, are valuable for ship sailing state analysis, maritime traffic flow prediction, and maritime accident analysis. However, due to the interferences of some factors, there sometimes exist errors in the collected AIS data, leading to the abnormity of AIS trajectory. Through in-depth analysis of a large amount of AIS data, this paper identifies several abnormal types of AIS trajectory. The automatic detection and restoration of trajectory anomalies are explored and an algorithm for that is developed with cubic spline interpolation model.

关 键 词：船舶自动识别系统 轨迹异常 自动检测 轨迹修复 

分 类 号：U675.7[交通运输工程—船舶及航道工程]