SIA OpenIR  > 光电信息技术研究室
Alternative TitleResearch on Target Tracking Algorithm and Real Time Implement Technology
Thesis Advisor罗海波
Keyword互信息 归一化互相关 Gabor小波 差值分解 Ti Dm642
Call NumberTP212/H87/2009
Degree Discipline模式识别与智能系统
Degree Name博士
Degree Grantor中国科学院沈阳自动化研究所
Place of Conferral沈阳
Abstract实时目标跟踪是模式识别、图像处理、计算机视觉、武器制导等领域的重要课题,而且在工业、军事和科学研究方面都具有广泛的应用。相关跟踪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。 同时,介绍了TI DM642硬件处理平台的基本构成和性能指标。作为正常算法分析的重要组成部分,对差值分解算法在该硬件处理平台上通过软件编程方式实现的复杂度进行了详细的分析与论证。 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。
Other AbstractReal-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. Correlation tracking technique has been widely used in the field of image traking, because it has high tracking precision. 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 scalling, 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 . At the same time, we introduce the struct and capability of TI DM642 development board.We analyze and argumentati on the complication of Difference Decomposition . Finaly, 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.
Contribution Rank1
Document Type学位论文
Recommended Citation
GB/T 7714
惠斌. 目标跟踪算法及其实时实现技术研究[D]. 沈阳. 中国科学院沈阳自动化研究所,2009.
Files in This Item:
File Name/Size DocType Version Access License
10001_14709T315惠斌_pa(3147KB) 开放获取--Application Full Text
Related Services
Recommend this item
Usage statistics
Export to Endnote
Google Scholar
Similar articles in Google Scholar
[惠斌]'s Articles
Baidu academic
Similar articles in Baidu academic
[惠斌]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[惠斌]'s Articles
Terms of Use
No data!
Social Bookmark/Share
All comments (0)
No comment.

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