摘要
提出一种新的鲁棒的旋转估计算法,以核相关滤波器理论为基础,通过在目标中心等角度间隔来采样一个样本金字塔,单独训练一个一维的角度估计滤波器,从而将目标旋转角估计问题变为一个检测问题。提出的角度估计方法具有通用性,可以辅助其它没有角度估计功能的跟踪器。在实验中,单独训练一个位移跟踪滤波器和一个尺度估计滤波器,结合本文提出的角度估计滤波器形成一个由三层滤波器组成的跟踪器。和经典的算法在不同测试数据上的对比实验表明,该算法能达到较高的跟踪精度。
A new robust rotation estimation method is proposed,which is based on kernelized correlation filters. A rotation pyramid is sampled around the target center and an extra 1D rotation filter is trained. Then the rotation estimation problem is tackled in a tracking-by-detection way. The rotation estimation approach is generic and can be incorporated with any tracking algorithm without a rotation estimation scheme. In the experiment,a three-layerbased tracker consisting of a 2D translation filter,a 1D scale filter and a 1D rotation filter is proposed trained and demonstrated to achieve more accurate results against state-of-the-arts on challenging dataset with rotation variation.
作者
李龙
王耀南
张振军
LI Long WANG Yao-nan ZHANG Zhen-jun(National Engineering Laboratory for Robot Vision Perception and Control Technologies, Changsha 410082, China College of Electrical and Information Engineering, Hunan University, Changsha 410082, China)
出处
《传感器与微系统》
CSCD
2017年第3期147-149,160,共4页
Transducer and Microsystem Technologies
基金
国家科技重大专项项目(07-Y30B10-9001-14/16)
国家自然科学基金资助项目(61501181)
关键词
目标跟踪
核相关滤波器
旋转估计
鲁棒跟踪
object tracking
kernelized correlation filters
rotation estimation
robust tracking
