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: 图像特征提取及目标匹配的理论和方法研究
其他题名: Research on the Methods of Feature Extracting and Target Matching
作者: 张志佳
导师: 史泽林 ; 黄莎白
分类号: TP391.4
关键词: 特征提取 ; 边缘特征匹配 ; 区域特征匹配 ; 代数特征匹配 ; 相关跟踪
索取号: TP391.4/Z36/2005
学位专业: 模式识别与智能系统
学位类别: 博士
答辩日期: 2006-01-15
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: 图像匹配作为图像处理及模式识别领域的一个基础问题,在光学制导、遥感测量、机器视觉、目标识别以及工业检测等领域都有广泛的应用。图像匹配的难点在于图像本身复杂多变、图像中存在运动目标以及各种成像过程中的畸变。寻找适应性强、精度高、计算速度快的匹配算法一直是当前的研究热点。本文以中国科学院知识创新工程项目“成像制导信息处理及平台控制新技术研究”为课题背景,结合实际成像制导项目要求,在对现有图像匹配方法深入分析的基础上,对边缘点特征、区域特征和代数特征进行了深入研究,给出了相应的特征提取方法和匹配算法。第一部分给出了图像匹配的数学模型并讨论了其发展现状。从特征空间、相似性测度、搜索策略这三个基本要素出发,阐述了图像匹配的基本过程及其影响因素。第二部分研究了预处理范畴上的图像去噪算法。提出了一种自适应确定图像傅立叶频域中条带成分的方法,从而有效去除图像中的条带噪声;结合基于奇异值分解的图像重构理论,建立了一个使最优去噪图像为能量函数最小解的能量模型,给出了一种针对高斯噪声的自适应图像去噪算法。第三部分研究了基于边缘点特征的匹配算法。针对边缘提取并二值化后的点特征图像,定义了一种新的距离函数以用于图像相关匹配。针对使用Hausdorff距离进行点集匹配时需要较大计算量这一缺点,提出了一种基于“搜索窗”的搜索算法来减少在不必要的位置上进行点集的Hausdorff匹配运算,从而提高了运算速度。第四部分研究了基于区域特征的匹配算法。提出了一种改进的基于势函数聚类的多阈值图像分割算法,以有效地提取图像的区域特征。在此基础上,对不同位置的目标区域,建立了一种不同隶属度的区域模板相关匹配方法并给出了实验结果。第五部分研究了基于代数特征的匹配算法。奇异值是对图像矩阵进行分解后提取出的一种代数特征,本文研究了奇异值分解与主成分分析之间的关系,证明了在正定对称矩阵条件下两者的等价性关系;提出了一种基于奇异值向量的目标匹配和跟踪算法;针对末制导过程中逼近目标时的尺度放大,定义了一个奇异值缩放不变特征,并据此提出了一种自适应相关匹配算法。
英文摘要: Image matching is an important and active branch of image processing and pattern recognition, and it has a wide application in fields such as missile guidance, remote sensing measurement, computer vision, target recognition, industry detection, etc. Several causes, such as complicated scene, moving target and many kinds of image distortion, make image matching difficult to implement. Research hotspots of image matching lie in looking for algorithms with strong adaptation, high precision and fast calculation speed. The dissertation is mainly aimed at image matching applications through three ways: edge feature, region feature and algebraic feature. And the dissertation designs different feature-extracting methods and image-matching algorithms corresponding to the three approaches respectively. All the researches in the dissertation have an imaging guidance application project background called “The research on new techniques for information processing and platform controlling in imaging guidance”. In the first section of the paper, the mathematical model of image matching and its research progress are illustrated. Based on three basic elements: feature domain, similar measurement and search strategy, the dissertation discusses the common process and influence factors of image matching. The second section deals with the de-noising algorithms for image pre-processing. An adaptive method is designed to obtain the stripe component in Fourier domain for image de-noising. And a Minimum Energy Model is given to remove image noises adaptively combined with the image reconstruction theory based on Singular Value Decomposition. In the third section, the edge-based image matching algorithms are studied. For the binary image after edge extracting, the paper defines a new distance function used in image correlation matching. Aiming at the high computation load when matching two point sets with Hausdorff Distance, this section also proposes a fast searching strategy named “searching window” to eliminate the matching positions that have little correlation to the template. In the fourth section, the region-based image matching algorithms are studied. An improved multi-threshold image segmentation algorithm based on potential function clustering is proposed to extract the region feature effectively. Then, the paper introduces a region template correlation matching algorithm which constitutes a method to measure the various contributing degree according to different regions. The corresponding results are presented in this section. In the fifth section, the algebraic feature-based image matching algorithms are studied. This section discusses the relation between Singular Value Decomposition and Principal Component Analysis, and gives a conclusion that the two methods are equivalent when they are applied to symmetric and positively definite matrices. Based on the characteristic of Singular Value, an algorithm for target matching and tracking is proposed. Aiming at the scale transformation of the approaching target, the paper proposes an adaptive correlation tracking algorithm based on the singular value invariant feature defined in the dissertation. Finally, future research is given in the end of the dissertation.
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/173321/9509
Appears in Collections:光电信息技术研究室_学位论文

Files in This Item:
File Name/ File Size Content Type Version Access License
图像特征提取及目标匹配的理论和方法研究.pdf(2393KB)----限制开放 联系获取全文

Recommended Citation:
张志佳.图像特征提取及目标匹配的理论和方法研究.[博士学位论文].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
[张志佳]'s Articles
CSDL cross search
Similar articles in CSDL Cross Search
[张志佳]‘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