摘要
基于全球船舶自动识别系统(Automatic Identification System,AIS)数据的船舶轨迹异常行为快速检测对于保障船舶航行安全、辅助安全监管具有重要意义。AIS数据具有容量大、更新频率快的特点,而当前AIS轨迹异常行为检测方法依赖于大量的训练样本与历史数据,实用性与普适性较差,难以用于船舶轨迹异常行为快速检测。为此,本文定义了船舶追踪、航速、航向、位置4种异常行为检测模型,提出了一种基于卡尔曼滤波的船舶AIS轨迹异常行为检测方法,实现了船舶AIS轨迹的异常行为快速检测与报警。实验选取经过我国东海部分地区3天的AIS数据,对实验结果的正确性与耗时进行分析,结果表明模型可以满足异常即时发现、即时处理的应用需求。
The rapid detection of abnormal ship trajectories of Automatic Identification System(AIS)data plays an important role in ensuring navigation safety of ships and assisting ship supervision.The AIS data has certain advantages such as of big volume and quick updation.However,current approaches of abnormal ship trajectory detection rely on big training samples and historical data,which make them difficult to be applied to real-life conditions.To solve this problem,we developed four kinds of models to define the abnormal ship trajectories by considering ship tracking,speed,heading and position properties.We also proposed a rapid abnormal ship trajectory detection method based on Kalman filter.The approach has been applied to the AIS data of 3 days in part of East China Sea.Results show that the approach is valid in fast detection of abnormal ship trajectory and real-time response of maritime transportation safety.
作者
杜志强
谭玉琪
仇林遥
DU Zhiqiang;TAN Yuqi;QIU Linyao(State Key Laboratory of Information Engineering in Surveying,Mapping and Remote Sensing,Wuhan 430079,China;China Academic of Electronics and Information Technology,Beijing 100041,China)
出处
《地理信息世界》
2021年第4期112-118,共7页
Geomatics World
基金
国家自然科学基金项目(41971347)
国家重点研发计划项目(2017YFC1404904)。
