一种改进的运动目标抗遮挡跟踪算法

赵龙; 肖军波

一种改进的运动目标抗遮挡跟踪算法

赵龙,
肖军波

北京航空航天大学飞行器控制一体化技术重点实验室, 北京 100191

基金项目: 国家自然科学基金重点资助项目(61039003 ); 中国航空基础科学基金资助项目(20100851018); 中国航天创新基金资助项目

详细信息

中图分类号: TP 391.4
计量
- 文章访问数: 1689
- HTML全文浏览量: 27
- PDF下载量: 1350
- 被引次数: 0
出版历程
- 收稿日期: 2012-03-02
- 网络出版日期: 2013-04-30

Improved algorithm of tracking moving objects under occlusions

Science and Technology on Aircraft Control Laboratory, Beijing University of Aeronautics and Astronautics, Beijing 100191, China

摘要

摘要: 为解决运动目标在遮挡情况下的跟踪问题,提出一种基于目标运动预测与自适应多子块模板匹配相结合的抗遮挡跟踪算法.该算法建立了多子块模板匹配相关算法中遮挡情况的判定、子块模板匹配及自适应更新等准则,采用卡尔曼滤波模型预测目标在遮挡时的运动轨迹,并利用一种基于目标速度矢量的模板定位规则实现目标在遮挡结束后的接力跟踪.将该算法应用于存在多种遮挡情况下的实际视频中进行测试,实验结果表明:该算法不仅能够实现在部分遮挡情况下的目标跟踪,而且能在严重遮挡、甚至完全遮挡情况下对刚体和非刚体目标进行稳定有效地跟踪,保持目标运动轨迹的可靠性和完整性.
- 计算机视觉 /
- 目标跟踪 /
- 模板匹配 /
- 卡尔曼滤波
Abstract: To better settle the problem of tracking objects under occlusions, an improved anti-occlusion tracking algorithm was proposed, based on the combination of motion prediction and adaptive multi-block template matching. The algorithm constituted the principle of detecting occlusions and the method of updating and matching multi-block templates. Kalman filter was applied to predict the object’s trajectory under occlusions. A template location rule based on the object’s velocity was adopted to re-track object after occlusion. Tracking experiments under various conditions of occlusion indicates that this algorithm succeeds to track rigid and non-rigid objects when heavily, even completely occluded, by static background. Moreover, the reliability and integrity of objects’ moving trajectories are guaranteed.
- computer vision /
- target tracking /
- template matching /
- Kalman filter

HTML全文

参考文献(12)

[1]	Lowe G D.Distinctive image features from scale-invariant key points[J].International Journal of Computer Vision,2004,60(2):91-110
[2]	Zhou Huiyu,Yuan Yuan,Shi Chunmei.Object tracking using SIFT features and mean shift[J].Computer Vision and Image Understanding,2009,113(3):345-352
[3]	Senior A,Hampapur A,Tian Yingli,et al.Appearance models for occlusion handling[J].Image and Vision Computing,2006,24(11):1233-1243
[4]	Zhang Jin,Sun Hongguang,Guan Weizhou,et al.Robust human tracking algorithm applied for occlusion handling //Proceedings International Conference on Frontier of Computer Science and Technology.Changchun: ,2010:546-551
[5]	Gross R,Mattews I,Baker S.Active appearance models with occlusion[J].Image and Vision Computing,2006,24(6):593-604
[6]	Khan M S,Shah M.Tracking multiple occluding people by localizing on multiple scene planes[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2009,31(3):505-518
[7]	Ablavsky V,Thangali A,Sclaroff S.Layered graphical models for tracking partially-occluded objects //Proceedings IEEE Conference on Computer Vision and Pattern Recognition.Alaska:IEEE,2008:1-8
[8]	Zhou Yue,Tao Hai.A background layer model for object tracking through occlusion //Proceedings IEEE Conference on Computer Vision and Pattern Recognition.Wisconsin:IEEE,2003:1079-1085
[9]	Wang Yan,Liu Tao,Li Ming.Object tracking with appearance-based kalman particle filter in presence of occlusions //Global Congress on Intelligent Systems.Hongkong:Conference Publishing Services,2009:288-293
[10]	Zhang Zui,Gunes H,Piccardi M.Tracking people in crowds by a part matching approach //Proceedings IEEE 5th International Conference on Advanced Video and Signal Based Surveillance.New Mexico:IEEE,2008:88-93
[11]	Latecki J L,Miezianko R.Object tracking with dynamic template update and occlusion detection //Proceedings IEEE 18th International Conference on Pattern Recognition.Hongkong:IEEE,2006:188-193
[12]	Pan Jiyan,Hu Bo.Robust occlusion handling in object tracking //Proceedings IEEE Conference on Computer Vision and Pattern Recognition.Minnesota:IEEE,2007:1-8