基于改进融合策略的多特征跟踪算法

doi:10.13973/j.cnki.robot.2016.0428

摘要
图/表
参考文献
相关文章 (15)

全文: PDF (1147 KB) HTML ( ) 输出: BibTeX \| EndNote (RIS)
引用本文:
夏瑜, 吴小俊, 李菊, 周立凡. 基于改进融合策略的多特征跟踪算法[J]. 机器人, 2016, 38(4): 428-436.DOI: 10.13973/j.cnki.robot.2016.0428. XIA Yu, WU Xiaojun, LI Ju, ZHOU Lifan. Multiple Features Tracking Algorithm Based on an Improved Fusion Strategy. ROBOT, 2016, 38(4): 428-436. DOI: 10.13973/j.cnki.robot.2016.0428.

摘要针对复杂场景中多特征融合视觉跟踪算法存在的问题，提出了一种基于改进融合策略的多特征跟踪算法．该算法在粒子滤波跟踪框架下通过改进融合方式，修正融合算法中常见的计算问题和量纲缺陷．在新的融合策略中引入目标状态关系图，可以提高跟踪的定位精度和稳定性．当目标状态存在偏差时，利用稀疏度对特征不确定性进行度量，在线选取最优状态，保证粒子状态空间质量，提高目标跟踪算法性能．实验结果表明，该算法比单特征跟踪、其他多特征融合策略跟踪算法具有更高的跟踪稳定性和更强的鲁棒性，3组实验的均方根误差小于1.2像素．

	服务

	把本文推荐给朋友
	加入我的书架
	加入引用管理器
	E-mail Alert
	RSS

关键词 ：视觉跟踪, 粒子滤波, 多特征融合, 稀疏度

Abstract：To solve the problem of visual tracking by multiple features fusion in complex scenes, a multiple features tracking algorithm based on an improved fusion strategy is presented. The proposed algorithm can solve the common computational problem and dimensional fault of the fusion algorithm by the improved fusion method under the particle filter framework. State relationship diagram introduced in the novel fusion strategy can improve the precision and robustness of tracking. When there are deviations of object state, this method can estimate the uncertainty measurement of features by sparsity and select the best state on line. In this way, the quality of state space can be guaranteed, and the performance of object tracking is improved. Experimental results show that the proposed method is stabler and more robust than the single feature algorithm and other multiple features fusion tracking algorithms. The root mean square errors in 3 experiments are all less than 1.2 pixels.

Key words： visual tracking particle filter multiple features fusion sparsity

收稿日期: 2016-03-01 录用日期: 2016-06-11

ZTFLH:

TP391.41

基金资助:教育部科学技术研究重大项目（311024）；国家自然科学基金（61373055，61300186，41501461）；江苏省自然科学基金（BK20140419）；江苏省高校自然科学研究项目（14KJB520001）；苏州大学高校省级重点实验室开放课题（KJS1522）

通讯作者: 夏瑜，cslgxiayu@163.com E-mail: cslgxiayu@163.com

作者简介: 夏瑜（1979-），男，博士，副教授．研究领域：计算机视觉，目标跟踪和数字图像处理．
吴小俊（1967-），男，教授，博士生导师．研究领域：模式识别，计算机视觉和计算智能．
李菊（1981-），女，博士，讲师．研究领域：运动目标检测与跟踪．

链接本文:

https://robot.sia.cn/CN/10.13973/j.cnki.robot.2016.0428 或 https://robot.sia.cn/CN/Y2016/V38/I4/428

[1] Smeulders A W M, Chu D M, Cucchiara R, et al. Visual tracking: An experimental survey[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2014, 36(7): 1442-1468.

[2] 张焕龙,胡士强,杨国胜.基于外观模型学习的视频目标跟踪方法综述[J].计算机研究与发展,2015,52(1):177-190.Zhang H L, Hu S Q, Yang G S. Video object tracking based on appearance models learning[J]. Journal of Computer Research and Development, 2015, 52(1): 177-190.
[3] Chu C T, Hwang J-N, Pai H-I, et al. Tracking human under occlusion based on adaptive multiple kernels with projected gradients[J]. IEEE Transactions on Multimedia, 2013, 15(7): 1602-1615.

[4] Gustafsson F, Gunnarsson F, Bergman N, et al. Particle filters for positioning, navigation and tracking[J]. IEEE Transactions on Signal Processing, 2002, 50(2): 425-437.

[5] Nummiaro K, Koller-Meier E B, van Gool L. An adaptive color-based particle filter[J]. Image and Vision Computing, 2003, 21(1): 99-110.

[6] Birchfield S. Elliptical head tracking using intensity gradients and color histograms[C]//IEEE Computer Society Conference on Computer Vision and Pattern Recognition. Los Alamitos, USA: IEEE Computer Society, 1998: 232-237.
[7] Wang J Q, Yagi Y. Integrating color and shape-texture features for adaptive real-time object tracking[J]. IEEE Transactions on Image Processing, 2008, 17(2): 235-240.

[8] Wang Q, Chen F, Xu W L. Adaptive multi-cue tracking by online appearance learning[J]. Neurocomputing, 2011, 74(6): 1035-1045.

[9] Erdem E, Dubuisson S, Bloch I. Fragments based tracking with adaptive cue integration[J]. Computer Vision and Image Understanding, 2012, 116(7): 827-841.

[10] Zoidi O, Nikolaidis N, Tefas A, et al. Stereo object tracking with fusion of texture, color and disparity information[J]. Signal Processing: Image Communication, 2014, 29(5): 573-589.

[11] Kittler J, Hatef M, Duin R P W, et al. On combining classifiers [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1998, 20(3): 226-239.

[12] 钟小品,薛建儒,郑南宁,等.基于融合策略自适应的多线索跟踪方法[J].电子与信息学报,2007,29(5):1017-1021.Zhong X P, Xue J R, Zheng N N, et al. An adaptive fusion strategy based multiple-cue tracking[J]. Journal of Electronics & Information Technology, 2007, 29(5): 1017-1021.
[13] 顾鑫,王海涛,汪凌峰,等.基于不确定性度量的多特征融合跟踪[J].自动化学报,2011,37(5):550-559.Gu X, Wang H T, Wang L F, et al.Fusing multiple features for object tracking based on uncertainty measurement[J]. Acta Automatica Sinica, 2011, 37(5): 550-559.
[14] Zhang X Q, Li W, Ye X Z, et al. Robust hand tracking via novel multi-cue integration[J]. Neurocomputing, 2015, 157(6): 296-305.
[15] Mei X, Ling H B. Robust visual tracking and vehicle classification via sparse representation[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2011, 33(11): 2259-2272.

[16] Wang D, Lu H C, Yang M H. Online object tracking with sparse prototypes[J]. IEEE Transactions on Image Processing, 2013, 22(1): 314-325.

[17] Zhuang B H, Lu H C, Xiao Z Y, et al. Visual tracking via discriminative sparse similarity map[J]. IEEE Transactions on Image Processing, 2014, 23(4): 1872-1881.

[18] Babu R V, Parate P, Acharya A K. Robust tracking with interest points: A sparse representation approach[J]. Image and Vision Computing, 2015, 33(1): 44-56.
[19] 章毓晋.图像工程[M].2 版.北京:清华大学出版社,2007:165-170.Zhang Y J. Image Engineering[M]. 2nd ed. Beijing: Tsinghua University Press, 2007: 165-170.
[20] 贾慧星,章毓晋.基于梯度方向直方图特征的多核跟踪[J].自动化学报,2009,35(10):1283-1289.Jia H X, Zhang Y J. Multiple kernels based object tracking using histograms of oriented gradients[J]. Acta Automatica Sinica, 2009, 35(10): 1283-1289.