抑制船尾拖纹的长时段船舶跟踪方法  

LONG-TERM SHIP TRACKING METHOD WITH SUPPRESSED STERN DRAG

作　　者：何水 陈黎[1,2] 陈姚节[1,2] He Shui;Chen Li;Chen Yaojie(Department of Computer Science and Technology,Wuhan University of Science and Technology,Wuhan 430081,Hubei,China;Hubei Province Key Laboratory of Intelligent Information Processing and Real-time Industrial System,Wuhan University of Science and Technology,Wuhan 430081,Hubei,China)

机构地区：[1]武汉科技大学计算机科学与技术学院,湖北武汉430081 [2]武汉科技大学湖北省智能信息处理与实时工业系统重点实验室,湖北武汉430081

出　　处：《计算机应用与软件》2024年第6期161-168,共8页Computer Applications and Software

基　　金：国家自然科学基金项目(61773297)。

摘　　要：船舶跟踪任务中水面拖纹干扰和目标遮挡会影响船舶跟踪的效果。提出一种抑制船尾拖纹的长时段船舶跟踪方法。对图像预处理,提升图像清晰度;使用离线训练的拖纹检测器检测水面拖纹,实时校正船舶跟踪框;根据感知哈希算法判断是否发生目标遮挡,遮挡时使用在线检测器对目标区域进行检测,找回船舶位置并初始化跟踪器,实现长时段跟踪。实验结果表明,该方法能有效抑制船尾拖纹对跟踪的影响,提高船舶跟踪精度,实现长时段船舶跟踪。In the ship tracking task,the surface dragging interference and target occlusion affect the result of ship tracking.Therefore,a long-term ship tracking method that suppresses stern drag is proposed.We pre-processed the image to improve image clarity.We used the offline training dragline detector to detect the location of the water surface dragline,and corrected the ship tracking frame in real time.The object occlusion was judged according to the perceptual Hash algorithm.We used an online detector to detect the target area,retrieved the ship's position and initialized the tracker to achieve long-term tracking.Experimental results prove that this method can effectively suppress the influence of stern drag on tracking,improve ship tracking accuracy,and achieve long-term ship tracking.

关 键 词：船舶跟踪 拖纹检测 孪生网络 长时段 重检测 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]