西安电子科技大学学报

• 研究论文 • 上一篇    下一篇

在线低秩表示的目标跟踪算法

王海军1,2;葛红娟1;张圣燕2   

  1. (1. 南京航空航天大学 民航学院,江苏 南京  210016;
    2. 滨州学院 山东省高校航空信息技术重点实验室,山东 滨州  256603)
  • 收稿日期:2015-09-07 出版日期:2016-10-20 发布日期:2016-12-02
  • 作者简介:王海军(1980-),男,南京航空航天大学博士研究生,E-mail: whjlym@163.com.
  • 基金资助:

    山东省自然科学基金资助项目(ZR2015FL009);滨州市科技发展计划资助项目(2013ZC0103);滨州学院科研基金资助项目(BZXYG1524)

Object tracking via online low rank representation

WANG Haijun1,2;GE Hongjuan1;ZHANG Shengyan2   

  1. (1. College of Civil Aviation, Nanjing Univ. of Aeronautics and Astronautics, Nanjing  210016, China;
    2. Key Lab. of Aviation Information Technology in Univ. of Shandong, Binzhou Univ., Binzhou  256603, China)
  • Received:2015-09-07 Online:2016-10-20 Published:2016-12-02

摘要:

针对传统的基于生成模式的跟踪方法对噪声及遮挡问题比较敏感,导致跟踪结果失败的问题,提出了以前几帧的跟踪结果作为观测矩阵,采用鲁棒的主元成分分析模型求解观测模型的低秩特征.当新的视频流到来时,不是把所有的跟踪结果矩阵作为观测矩阵.并提出了新的增量鲁棒的主元成分分析模型,采用增广拉格朗日算法求解新矩阵的低秩特征,并以此低秩矩阵在贝叶斯框架下建立跟踪模型,用恢复的低秩特征更新字典矩阵.将文中方法与其他6种跟踪算法在8种跟踪视频上进行跟踪对比.实验结果表明,所提出的方法具有较低的像素中心位置误差和较高的重叠率.

关键词: 目标跟踪, 低秩特征, 鲁棒的主成分分析模型, 字典矩阵

Abstract:

Object tracking is an active research topic in computer vision. The traditional tracking methods based on the generative model are sensitive to noise and occlusion, which leads to the failure of tracking results. In order to solve this problem, the tracking results of the first few frames are used as the observation matrix, and the low rank features of the observation model are solved by the the RPCA model. When the new video streams come, a new incremental RPCA is proposed to compute the new observation matrix by the augmented Lagrangian algorithm. The tracking model is established in the Bayesian framework, and the dictionary matrix is updated with the low rank feature. We have tested the proposed algorithm and six state-of-the-art approaches on eight publicly available sequences. Experimental results show that the proposed method has a lower pixel center position error and a higher overlap ratio.

Key words: object tracking, low rank feature, RPCA model, dictionary matrix