基于中心点检测与卡尔曼滤波的船舶跟踪算法  

Ship Tracking Algorithm Based on CenterNet Detection and Kalman Filter

作　　者：熊超 周海峰[1,2] 郑东强 林忠华[3] 张兴杰[4] 关天敏[5] XIONG Chao;ZHOU Hai-feng;ZHENG Dong-qiang;LIN Zhong-hua;ZHANG Xing-jie;GUAN Tian-min(College of Marine Engineering,Jimei University,Xiamen Fujian 361021,China;Key Laboratory of Naval Architecture and Ocean Marine Engineering of Fujian Province,Xiamen Fujian 361021,China;School of Marine Equipment and Mechanical Engineering,Jimei University,Xiamen Fujian 361021,China;Nautical Institute,Jimei University,Xiamen Fujian 361021,China;College of Ocean Information Engineering,Jimei University,Xiamen Fujian 361021,China)

机构地区：[1]集美大学轮机工程学院,福建厦门361021 [2]福建省船舶与海洋工程重点实验室,福建厦门361021 [3]集美大学海洋装备与机械工程学院,福建厦门361021 [4]集美大学航海学院,福建厦门361021 [5]集美大学海洋信息工程学院,福建厦门361021

出　　处：《广州航海学院学报》2022年第2期32-37,共6页Journal of Guangzhou Maritime University

基　　金：福建省自然科学基金项目(2021J01839,2018J01495);产学研项目(S20127);福建省教育厅项目(JAT200242,JAT170318)。

摘　　要：为了提升船舶跟踪效率,提出一种基于中心点检测与卡尔曼滤波的船舶目标跟踪算法.该算法根据经典的Tracking-by-Detection模式,采用中心点算法检测船舶目标,利用卡尔曼滤波算法为检测目标创建跟踪器,并通过匈牙利算法解决检测目标与跟踪器预测目标之间的匹配关联问题,从而实现对实际场景视频中船舶目标的稳健跟踪.实验结果表明,基于中心点检测与卡尔曼滤波的跟踪算法能够有效应对船舶部分遮挡、船舶形变、光线不足、小目标等复杂情况.In order to improve the ship tracking efficiency,a ship target tracking algorithm based on CenterNet detection and Kalman filter is proposed.According to the classic Tracking-by-Detection mode,the algorithm uses the CenterNet algorithm to detect the ship target,uses the Kalman filter algorithm to create a tracker for the detected target,and uses the Hungarian algorithm to solve the matching problem between the detected target and the target predicted by the tracker,so as to achieve robust tracking of ship targets in actual scene videos.The experimental results show that the ship tracking algorithm based on CenterNet detection and Kalman filter can deal with complex situations such as partial occlusion of ships,ship deformation,insufficient light,and small targets.

关 键 词：中心点检测 卡尔曼滤波算法 匈牙利算法 船舶目标跟踪 

分 类 号：TN911[电子电信—通信与信息系统]