首页 > 论文 > 中国光学 > 2卷 > 5期(pp:388-393)

遮挡情况下目标跟踪算法综述

Review of tracking algorithms under occlusions

  • 摘要
  • 论文信息
  • 参考文献
  • 被引情况
  • PDF全文
分享:

摘要

目标跟踪过程中出现的遮挡问题一直是目标跟踪中的难点。如何有效处理遮挡,尤其是严重遮挡和全部遮挡是评价目标跟踪算法鲁棒性的关键指标。本文对国内外现有的遮挡情况下的目标跟踪算法进行了总结和分类,分别介绍了基于中心加权、子块匹配、轨迹预测、贝叶斯理论的多种遮挡目标跟踪算法,特别描述了综合多种算法优势的多算法融合跟踪算法。阐述了各种算法的基本思想及其遮挡处理功能,分析了各种算法的优缺点,并指出了它们的适用场合。文章最后提出了遮挡情况下目标跟踪算法存在的问题,特别指出将多传感器融合技术用于有遮挡的目标跟踪是该项技术的发展方向之一。

Abstract

The occlusions occurred in the target tracking is one of the difficulties in image processing, which has become a crucial factor of the robustness of tracking algorithm to deal with the occlasions effectively in tracking, especally severe occlusion and total occlusion. This paper summarizes and classifies the main tracking algorithms under occlusions commonly used at home and abroad, and introduces several kinds of popular tracking algorithms under occlusions based on central weights, part matching, track prediction, and Bayesian theory, respectively. Especially, it describes a fusion algorithm combined different advantages of some algorithms. It also gives the basic ideas of these algorithms and their processing abilities, and analyzes the advantages and disadvantages of these algorithms. Finally, it focus on the problems and points out that the multi-sensor fusion technology will be conductive to the improvement of the tracking algorithms under occlusions.

中国激光微信矩阵
补充资料

中图分类号:TP301.6;TP391

所属栏目:综述

收稿日期:2009-06-12

修改稿日期:2009-08-16

网络出版日期:0001-01-01

作者单位    点击查看

薛陈:中国科学院 长春光学精密机械与物理研究所,吉林 长春 130033中国科学院 研究生院,北京 100039
朱明:中国科学院 长春光学精密机械与物理研究所,吉林 长春 130033
刘春香:长春理工大学 电子信息工程学院,吉林 长春 130022

联系人作者:薛陈(achen5225@sina.com)

备注:薛陈(1983—),男,四川成都人,博士研究生,主要从事图像处理、目标跟踪等方面的研究。

【1】COMANICIU D,RAMESH V,MEER P. Kernel-based object tracking[J]. IEEE T. Pattern Anal.,2003:564-577.

【2】江泽涛,赵榕春,黎明.一种基于相关的分层匹配与目标跟踪算法[J].航空学报,2006,27(4):670-675.
JIANG Z T,ZHAO R CH,LI M. A correlation based layered matching and target tracking method[J]. Acta Aeronautica et Astronautics Sinica,2006,27(4):670-675.(in Chinese)

【3】王明波,周亚凡,刘颖.遮挡情况下目标跟踪算法研究[J].现代雷达,2008,30(7):52-55.
WANG M B,ZHOU Y F,LIU Y. A study on tracking method for moving target under screening[J]. Modern Radar,2008,30(7):52-55(in Chinese).

【4】JEYAKAR J,VENKATESH B R,RAMAKRISHNAN K R. Robust object tracking using local kernels and background information[C]. IEEE Int. Conf. on Image Proc.,2007:49-52.

【5】JEYAKAR J,VENKATESH B R,RAMAKRISHNAN K R. Robust object tracking with background-weighted local kernels[J]. Comput. Vis. Image Und.,2008:1-14.

【6】WEN ZH Q,CAI Z X. A robust object racking approach using Mean Shift[C]. 3th Int. Conf. on Nat. Computat.(ICNC 2007),2007.

【7】MAGGIO E,CAVALLARO A. Multi-part target representation for color tracking[C]. ICIP 2005. IEEE Int. Conf. on Image Proc.,2005,1:729-732.

【8】WANG F L,YU SH Y,YANG J. A novel fragments-based tracking algorithm using Mean Shift[C]. 2008 10th Intl. Conf. on Control, Automation, Robotics and Vision, Hanoi,Vietnam,2008:694-698.

【9】CAULFIELD D,DAWSON-HOWE K. Evaluation of multi-part models for Mean-Shift tracking[C]. Int. Machine Vision and Image Proc. Conf.,2008,77-82.

【10】ZHANG Z,GUNES H,PICCARDI M. Tracking people in crowds by a part matching approach[C]. IEEE 5th Int. Conf. on Advanced Video and Signal Based Surveillance,2008:88-93.

【11】孙中森,孙俊喜,宋建中.一种分块表示的彩色目标跟踪算法[J].电子应用技术,2007,3:55-57.
SUN ZH S,SUN J X,SONG J ZH. A color target tracking algorithm based on multi-part representation[J]. Appl. Electronic Technique,2007,3:55-57.(in Chinese).

【12】常发亮,马丽,乔谊正.遮挡情况下基于特征相关匹配的目标跟踪算法[J].中国图像图形学报,2006,11(6):877-882.
CHANG F L,MA L,QIAO Y ZH. Target tracking algorithm under occlusion based on correlation matching[J]. J. Image and Graphics,2006,11(6):877-882.(in Chinese)

【13】常发亮,马丽,乔谊正.遮挡情况下的视觉目标跟踪方法研究[J].控制与决策,2006,21(5):503-507.
CHANG F L,MA L,QIAO Y ZH. Study on vision target tracking under occlusion[J]. Control and Decision,2006,21(5):503-507.(in Chinese)

【14】ADAM A,RIVLIN E,SHIMSHONI I. Robust fragments-based tracing using integral histogram[C]. Proc. of the IEEE Society Conf. on Comput. Vis. and Pattern Recognition,2006:798-805.

【15】KALMAN R E. A new approach to linear filtering and prediction problems[J]. Trans. ASME-J. of Basic Eng.,1960,82(Series D):35-45.

【16】WANG ZH Q,FAN Y F,ZHANG G L,et al.. Robust face tracking algorithm with occlusions[J]. SPIE,2007,6786:67861X.

【17】RIBEIRO M I. Kalman and extended Kalman filter:concept, derivation and properties[R/OL]. Institute for Systems and Robotics,2004[2009-01-11].http://user.isr.ist.utl.pt/~mir/pub/Kalman.pdf.

【18】JULIER S,UHLMANN J. A new extension of the Kalman filter to nonlinear systems[J]. SPIE,1997,3068:182-193.

【19】GORDON N J,SALMOND D J,SMITH A F M. Novel approach to nonlinear/non-gaussian bayesian state estimation[J]. IEEE Proc-F,1993,140(2):107-113.

【20】DOUCET A,GODSILL S J,ANDRIEU C. On sequential monte carlo sampling methods for bayesian filtering[J]. Stat. Compu.,2000,10(3):197-208.

【21】ISARD M,BLAKE A. CONDENSATION-Conditional density propagation for visual tracking[J]. Int. J. Comput. Vis.,1998,29(1):5-28.

【22】孟勃.局部最优粒子滤波目标跟踪算法的研究和应用[D].北京:中国科学院研究生院,2008.
MENG B. Research and application of the local optimal particle filter target tracking algorithm[D]. Beijing:Graduate University of Chinese Academy of Science,2008.(in Chinese)

【23】CORVEE E,VELASTIN S,JONES G A. Occlusion tolerent tracking using hybrid prediction schemes[J]. Acta Automation Sinica,2003,29(3):356-349.

【24】陈爱华,孟勃,朱明.多模式融合的目标跟踪算法[J].光学 精密工程,2009,17(1):185-190.
CHEN A H,MENG B,ZHU M. Multi-pattern fusion algorithm for target tracking[J]. Opt. Precision Eng.,2009,17(1):185-190.(in Chinese)

【25】陈爱华.复杂环境下多模式融合的视频跟踪算法研究[D].北京:中国科学院研究生院,2009.
CHEN A H. Research on multi-mode fusion algorithm for visual target tracking under complex enviroment[D]. Beijing:Graduate University of Chinese Academy of Science,2009.(in Chinese)

【26】LOWE D G. Object recognition from local scale-invariant features[C]. Proc. Int. Conf. on Comp. Vis.(ICCV''1999),1999:20-27.

【27】LOWE D G. Distinctive image features from scale-invariant keypoints[J]. Int. J. Comput. Vis.,2004,60(2):99-110.

您的浏览器不支持PDF插件,请使用最新的(Chrome/Fire Fox等)浏览器.或者您还可以点击此处下载该论文PDF