改进的核相关滤波算法在自航模动态目标跟踪应用  

Application of improved kernel correlation filtering algorithm in small ship dynamic target tracking

作　　者：程子一 刘志林[1] CHENG Ziyi;LIU Zhilin(College of Automation,Harbin Engineering University,Harbin 150001,China)

机构地区：[1]哈尔滨工程大学自动化学院,黑龙江哈尔滨150001

出　　处：《应用科技》2019年第1期36-42,共7页Applied Science and Technology

基　　金：国家自然科学基金项目(51379044);中央高校基本科研业务项目(HEUCFG2018)

摘　　要：针对核相关滤波(KCF)算法在尺度变换和严重遮挡造成跟踪失败的缺点,本设计在KCF算法基础上通过增加尺度变换框与跟踪效果检测的方法,对KCF算法进行了改进,解决了在目标被遮挡或者离开屏幕跟踪失败的问题。实验表明,该方法应用在自航模的动态目标跟踪上,使自航模在目标大小变化、目标被遮挡和逃离视野时,能成功地找回目标并继续跟踪。Kernel correlation filtering(KCF)algorithm is a target tracking algorithm,or KCF algorithm.KCF algorithm has shortcomings in scale transformation and severe occlusion,which causes failure of tracking.The KCF algorithm is improved by increasing the scale transform frame and the method for detection of tracking effect,which solves the problem of failure of the target being blocked or leaving the screen tracking.The experiment shows that when the method is applied to the dynamic target tracking of a small ship,it can make the small ship retrieve the target and continue to track the target when the target has changed in size,obscured and escaped from the field of vision.

关 键 词：目标跟踪 相关滤波 动态目标 机器视觉 尺度变换 机器学习 船舶控制 循环矩阵 

分 类 号：TP249[自动化与计算机技术—检测技术与自动化装置]