运动状态与尺度估计的核相关目标跟踪方法

doi:10.11896/j.issn.1002-137X.2017.11A.040

摘要/Abstract

摘要： 视觉跟踪在视频智能监控和机器人等领域有着广泛应用。基于相关滤波分类器,提出了具有运动状态估计和目标尺度估计的视觉目标跟踪方法。该方法将粒子滤波与核相关滤波方法相结合,首先估算运动目标的位置,然后执行尺度相关滤波器来估算目标的尺度,以使算法对尺度变化的运动目标具有更强的适应能力。该方法在传统的KCF跟踪算法的基础上引入了一种基于概率的运动状态估计方法,可以获得更加稳定的目标信息,并减少背景干扰信息的引入,从而在复杂场景下具有更强的抗干扰性。使用benchmark数据集对所提方法进行了测试实验,并和其他已有的若干视觉跟踪方法进行了对比实验,结果验证了所提算法的高效性,且所提方法在目标尺度变化、光照变化、姿态变化、部分遮挡、旋转及快速运动等复杂情况下均有较强的适应性。

关键词: 目标跟踪,机器视觉,相关滤波器,运动状态估计,尺度空间估计

Abstract: Visual tracking has a wide range of applications in various fields such as video intelligent monitoring and eyes of robot.Based on discriminatory correlation filter,a visual target tracking method with motion estimation and scale estimation method was proposed.First,this method combines translation kernel correlation filter and particle filter method to estimate the position of moving targets on the frame.And then,it executes the scale correlation filter to be stronger ability to adapt scale change of moving object.In traditional KCF tracking algorithm,this method introduces a motion state estimation method based on probability,which can obtain more stable signal of target,and reduces the introduction ofthe background interference information at the same time,leading to it has stronger anti-jamming in complex scenarios.On benchmark data set,we took experiment to test this method,and compared our method with the current advanced tracking methods to verify the efficiency of the proposed algorithm in this paper.It has strong adaptability under complex conditions,the change of scale,illumination,change of pose,partial sheltering,rotating and rapid movement etc.

Key words: Target tracking,Machine vision,Correlation filter,Motion state estimation,Scale estimation

朱航江,朱帆,潘振福,朱永利. 运动状态与尺度估计的核相关目标跟踪方法[J]. 计算机科学, 2017, 44(Z11): 193-198. https://doi.org/10.11896/j.issn.1002-137X.2017.11A.040

ZHU Hang-jiang, ZHU Fan, PAN Zhen-fu and ZHU Yong-li. Visual Object Tracking Method with Motion Estimation and Scale Estimation[J]. Computer Science, 2017, 44(Z11): 193-198. https://doi.org/10.11896/j.issn.1002-137X.2017.11A.040

参考文献

[1] YILMAZ B A,et al.Object Tracking:A Survey [J].ACM Journal of Computing Surveys,2006,38(4):13.
[2] CANNONS K.A review of visual tracking[J].Dept.comput.sci.eng,2008:42.
[3] ZHANG K,ZHANG L,LIU Q,et al.Fast Visual tracking viadese spatio-temporal context learning [C]∥Proc.of the European Conference on Computer Vision.2014.
[4] PETROSINO A,KRISTAN M,MATAS J,et al.The visual object tracking vot2015 challenge results[C]∥Computer Vision-ICCV.2016:564-586.
[5] CASASENT D.Unified synthetic discriminant function computational formulation [J].Applied Optics,1984,23(10):1620-1627.
[6] HESTER C F,CASASENT D.Multivariant technique for multiclass pattern recognition [J].Applied Optics,1980,19(11):1758-1761.
[7] KRISTAN M,PFLUGFELDER R,Leonardis A,et al.The Visual Object Tracking VOT2014 challenge results[J].Lecture Notes in Computer Science,2014,8926:191-217.
[8] BOLME D S,BEVERIDGE J R,DRAPER B,et al.Visual object tracking using adaptive correlation filters[C]∥2010 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).IEEE,2010:2544-2550.
[9] HENRIQUES J F,RUI C,MARTINS P,et al.Exploiting theCirculant Structure of Tracking-by-Detection with Kernels [J].Lecture Notes in Computer Science,2012,7575(1):702-715.
[10] DANELLJAN M,KHAN F S,FELSBERG M,et al.AdaptiveColor Attributes for Real-Time Visual Tracking[C]∥2014 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).IEEE,2014:1090-1097.
[11] HENRIQUES J F,CASEIRO R,MARTINS P,et al.High-Speed Tracking with Kernelized Correlation Filters [J].IEEE Transactions on Pattern Analysis & Machine Intelligence,2014,37(3):1.
[12] DANELLJAN M,HGER G,KHAN F,et al.Accurate scaleestimation for robust visual tracking[C]∥British Machine Vision Conference.Nottingham,BMVA Press,2014.
[13] NUMMIARO K,KOLLER-MEIER E,G OOL L V.An adaptive color-based particle filter[J].Image & Vision Computing,2003,21(1):99-110.
[14] LU H,JIA X,YANG M H.Visual tracking via adaptive structural local sparse appearance model[C]∥IEEE Conference on Computer Vision & Pattern Recognition.IEEE Computer Society,2012:1822-1829.
[15] DONG W,LU H,YANG M H.Online object tracking withsparse prototypes [J].IEEE Transactions on Image Processing,2013,22(1):314-325.
[16] WU Y,LIM J,YANG M H.Object Tracking Benchmark [J].IEEE Transactions on Pattern Analysis & Machine Intelligence,2015,37(9):1834-1848.
[17] ZHONG W.Robust object tracking via sparsity-based collaborative model [C]∥IEEE Conference on Computer Vision & Pattern Recognition.2012:1838-1845.
[18] SCHLKOPF B,SMOLA A.Learning With Kernels:SupportVector Machines,Regularization,Optimization,and Beyond [J].Journal of the American Statistical Association,2005,16(3):781-781.
[19] RIFKIN R,YEO G,POGGIO T.Regularized Least-SquaresClassification [J].Acta Electronica Sinica,2003,190(1):93-104.
[20] GRAY R M.Toeplitz and circulant matrices:a review[J].Foundations & Trends@ in Communications & Information Theory,2005,2(3):155-239.
[21] 张华军,赵金,罗慧.基于拟牛顿法的同时扰动随机逼近算法[J].华中科技大学学报(自然科学版),2014(9):1-4.
[22] 王法胜,赵清杰.一种用于解决非线性滤波问题的新型粒子滤波算法[J].计算机学报,2008,31(2):346-352.
[23] WU Y,LIM J,YANG M H.Online Object Tracking:A Benchmark[C]∥2013 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).IEEE,2013:2411-2418.
[24] BABENKO B,YANG M H,BELONGIE S.Robust ObjectTracking with Online Multiple Instance Learning[J].IEEE Transactions on Pattern Analysis & Machine Intelligence,2011,33(8):1619-1632.

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed