基于群运动特征的渔网示位标智能识别算法研究  

Research on intelligent identification algorithm of fishing net indicating markers based on group motion characteristics

作　　者：郭小飞 刘家兵 胡青[2] GUO Xiaofei;LIU Jiabing;HU Qing(Northern Navigation Service Center,Maritime Safety Administration,People's Republic of China,Tianjin 300000,China;Information Science and Technology College,Dalian Maritime University,Dalian 116026,China)

机构地区：[1]交通运输部北海航海保障中心,天津300000 [2]大连海事大学信息科学技术学院,辽宁大连116000

出　　处：《中国航海》2024年第3期1-9,20,共10页Navigation of China

基　　金：国家重点研发计划课题(2021YFB3901502);辽宁省科学技术计划项目(2021JH1/10400008)。

摘　　要：随着海上船舶数量的增长,复杂的海上业务使海上通信链路负载和信道占用率明显增加,信道拥挤程度加剧。而海上通信系统相互独立,各系统之间存在管理不统一和带宽分配不合理等问题。因此,国际航标协会(IALA)指出,当海上信道负载超过50%时,信息阻塞等问题将造成信号传输滞后。为保护船舶自动识别系统(AIS)信道资源,保障海上交通安全,针对渔网示位标信息不规范造成识别困难,进而影响船舶航行安全的问题,提出一种基于群运动特征的渔网示位标智能识别算法。该算法基于渔网示位标群运动特征模型,结合监督和非监督机器学习算法,实现在海量船舶和渔网示位标混杂的AIS数据中智能识别渔网示位标。结果表明:针对随机海域内的AIS数据,该算法可有效识别渔网示位标,准确率超95%,有效提升对渔网示位标的管控能力,降低海上航行安全潜在风险。The growing number of ships at sea and the complexity of maritime operations significantly increase the load on maritime communication links and channel occupancy,intensifying channel congestion.Maritime communication systems often operate independently,leading to inconsistent management and inefficient bandwidth allocation.IALA(The International Association of Marine Aids to Navigation and Lighthouse Authorities)warns that when maritime channel load exceeds 50%,signal transmission may experience delays due to information blockages.To protect AIS(Automatic Identification System)channel resources and ensure maritime traffic safety,I propose an intelligent identification algorithm that addresses the challenge of recognizing fishing net markers due to irregular information,which impacts navigation safety.This algorithm models the group movement characteristics of fishing net markers and combines supervised and unsupervised learning methods to accurately identify fishing net markers within large AIS data sets that include ships and fishing nets.The results show that the algorithm effectively identifies fishing net markers,achieving an accuracy rate of over 95%,enhancing control over these markers and reducing potential risks to maritime safety.

关 键 词：船舶自动识别系统 渔网示位标 群运动 智能识别 

分 类 号：U675.7[交通运输工程—船舶及航道工程]