中国科学院沈阳自动化研究所机构知识库
Advanced  
SIA OpenIR  > 光电信息技术研究室  > 学位论文
题名: 差值分解方法及其在目标跟踪中的应用
作者: 闫占德
导师: 罗海波
关键词: 相关跟踪 ; 差值分解 ; 遮挡鲁棒性 ; 光照补偿 ; 自适应
学位专业: 模式识别与智能系统
学位类别: 硕士
答辩日期: 2006-05-31
授予单位: 中国科学院沈阳自动化研究所
学位授予地点: 中国科学院沈阳自动化研究所
作者部门: 光电信息技术研究室
中文摘要: 目标跟踪是模式识别、图像处理、计算机视觉领域的重要课题,而且在工业、军事和科学研究方面都具有广泛的应用,例如虚拟现实环境、模拟训练、自动导航等。目标跟踪算法的优劣直接影响着跟踪的稳定性和精确性。相关跟踪由于具有精度高、跟踪距离远、可靠性高、较强的局部抗干扰能力等优点,是目前使用最广泛的跟踪算法。但传统相关跟踪方法以假设目标仅发生平移运动为前提,当目标仅发生平移时能够获得理想的跟踪效果。但当目标尺度和灰度变化时,这种算法往往表现出一定的不适应性。差值分解(Difference Decomposition)最早于1997年Michael Gleicher提出,并被应用于目标跟踪,图像配准等领域。由于具有计算速度快,对目标变化适应性好等特点,被认为是目标跟踪中的一种有效的方法。本文在跟踪算法中引入了这种方法,力图解决传统相关跟踪所出现的上述问题。在研究差值分解(Difference Decomposition)理论的同时,对使用该方法在实际应用中遇到的问题进行了深入的分析和大量的实验。主要包括:算法应用中一些参数的选择对算法的影响,算法迭代中参数更新的方法等。并在应用中发现了算法的不足之处,提出了相应的改进方法。主要的改进包括两个方面:(1) 针对算法遇到光照引起的灰度线性变化而出现的跟踪失败问题,采用的解决办法是:用两个可以直接得到的图像矩阵对原模板图像进行光照补偿,在迭代中逐步优化光照补偿的系数,进而达到消除光照影响的目的。(2) 针对算法在遇到有外界物体遮挡候跟踪效果下降的问题,提出利用M估计函数对模板图像与目标图像的差值图像各象素点赋于不同权值,并通过迭代循环逐步来更新权值,进而达到消除遮挡干扰的目的。最后本文构建了一个比较完整的跟踪流程,将改进后的跟踪算法应用到所建立的跟踪流程中。采用Matlab工具对算法进行了开发,并使用序列图像对算法进行了跟踪仿真实验,为算法将来的实际应用奠定了良好的基础。
英文摘要: Real-time object tracking is an important subject in computer vision, pattern recognition, image processing and artificial intelligence. It is widely applied in the areas of industry, military and science research, such as automatically navigation, virtual reality, video-conference. The arithmetic used has directly impact on the stability and precision of target tracking. Correlation tracking technique has been widely used in the field of image based guiding weapon system. This method imply a presupposition that there exist only translational target motions, thus this traditional tracking method is hold provided that only translation shift is occurred on the target, and is not capable of adapting well to scaling, rotational changes of target which happen in most applications. Difference Decomposition was first proposed by Michael Gleicher in 1997 and has been wildly used in the area of target tracking and image registration. Difference Decomposition has been regarded as a kind of effective target tracking methods because it has fast speed and good adaptability to target variations. This tracking arithmetic was used in this thesis, for the sake of solving the disadvantage of Correlation tracking mentioned above. Our research follows the theory of Difference Decomposition, and has analysed the problems in detail which will be met in application: The impact of the parameters have on the arithmetic performance; How to update the parameter in iterations. Also we have found some disadvantages of this arithmetic in application, then proposes the improved method . We have improved the technology of target tracking using Difference Decomposition in three aspects: 1) In order to solve the problem of tracking fail caused by illumination, we use two images that can easily get as basis image for the illumination compensation of the template. The parameters of illumination compensation is optimized in iterations step by step, and the impact of illumination is to be eliminated. 2) For the sake of solving the problem that tracking effect becomes worse when some other object occlude, each point is endowed with a weight value that is estimated with “M-estimator”. Then the value is updated each iteration, and at last the disturbance is to be eliminated. Finally, we have developed a system of target tracking using the improve method which we proposed in this paper. Through designing each function module rationally and adopting Matlab tool to carry on system development. Tracking experiment of sequence image is also done with this system. The system developed in this paper has settled good foundation for developing the application software in future.
语种: 中文
产权排序: 1
内容类型: 学位论文
URI标识: http://ir.sia.cn/handle/173321/9100
Appears in Collections:光电信息技术研究室_学位论文

Files in This Item:

There are no files associated with this item.


Recommended Citation:
闫占德.差值分解方法及其在目标跟踪中的应用.[硕士学位论文].中国科学院沈阳自动化研究所.2006
Service
Recommend this item
Sava as my favorate item
Show this item's statistics
Export Endnote File
Google Scholar
Similar articles in Google Scholar
[闫占德]'s Articles
CSDL cross search
Similar articles in CSDL Cross Search
[闫占德]‘s Articles
Related Copyright Policies
Null
Social Bookmarking
Add to CiteULike Add to Connotea Add to Del.icio.us Add to Digg Add to Reddit
所有评论 (0)
暂无评论
 
评注功能仅针对注册用户开放,请您登录
您对该条目有什么异议,请填写以下表单,管理员会尽快联系您。
内 容:
Email:  *
单位:
验证码:   刷新
您在IR的使用过程中有什么好的想法或者建议可以反馈给我们。
标 题:
 *
内 容:
Email:  *
验证码:   刷新

Items in IR are protected by copyright, with all rights reserved, unless otherwise indicated.

 

 

Valid XHTML 1.0!
Copyright © 2007-2016  中国科学院沈阳自动化研究所 - Feedback
Powered by CSpace