中国科学院光电技术研究所机构知识库
Advanced  
IOE OpenIR  > 光电技术研究所博硕士论文  > 学位论文
题名:
成像跟踪系统中的盲图像复原研究
作者: 张建林
学位类别: 博士
答辩日期: 2008-06-10
授予单位: 中国科学院光电技术研究所
授予地点: 光电技术研究所
导师: 张启蘅
关键词: 盲图像复原 ; 盲解卷积 ; 逆问题 ; 图像处理
其他题名: The study of blind image restoration in imaging and tracking system
学位专业: 信号与信息处理
中文摘要: 由于大气湍流扰动的影响,成像跟踪系统中目标的成像发生漂移、闪烁和模糊等降质现象,从而对目标的探测、识别和跟踪等带来了极大的困难。尤其是面临各领域对成像跟踪系统的探测能力、定位精度和跟踪精度的要求越来越高的紧迫需求,大气湍流对系统成像质量的影响显得更为突出,而克服大气湍流的影响提高系统成像的质量则成为系统急需解决的关键问题。本文紧密围绕克服大气湍流的影响、改善成像跟踪系统在大气湍流干扰下的成像质量这一主题而展开,进行大气湍流条件下降质图像的盲复原研究,旨在采用盲图像复原技术改善系统成像质量、提升系统性能。本文重点研究了盲图像复原的相关理论,提出了新的盲图像复原算法。 盲图像复原是指在未知或不完全确知目标图像与成像系统点扩展函数(又称为退化过程函数)的情形下,直接利用系统观测所得的降质图像寻求目标图像的最佳估计,以削弱或去除退化过程对目标图像的降质,进而改善目标图像质量。盲图像复原因无需系统点扩展函数知识,使其适宜于克服大气湍流的影响。但盲图像复原由于其可利用信息很少,其求解非常困难。本文结合大多数成像系统的点扩展函数(包括大气湍流长曝光成像的系统点扩展函数)都是G类函数的特点,针对大气湍流长曝光降质图像及其它系统点扩展函数为G类函数的降质图像的盲复原,提出基于DAPEX的盲图像复原算法,该算法通过将图像功率谱特征模型引入盲图像复原,实现直接从降质图像进行系统点扩展函数的估计。同时,针对盲图像复原在复原过程放大噪声的问题,本文在基于DAPEX的盲图像复原算法的基础上,提出基于预去噪和DAPEX的盲图像复原算法,实现复原图像的同时有效抑制噪声的放大。 由于大气湍流的瞬时运动非常随机和复杂,因而导致大气湍流短曝光成像的系统点扩展函数难以描述,这使得针对大气湍流短曝光降质图像的盲复原不能对系统点扩展函数类型作任何假设而变得更为困难。针对该问题,本文提出基于乘性迭代的盲图像复原算法,该算法由于在交替迭代估计中采用了乘积的形式,确保了在解的初始估计非负的情形下,解在整个迭代过程中保持非负,有效避免了算法通过引入非负惩罚约束项或对解进行强制非负约束等所带来的计算复杂性,同时该算法的乘积形式避免了经典的图像复原算法如RLA(Richardson-Lucy Algorithm)和图像空间重构算法ISRA (Image Space Reconstruction Algorithm)中的除运算所带来的数值计算的不稳定。值得指出的是该算法能实现单帧降质图像的高分辨率复原,给出了一种新的盲图像复原解决途径。 针对大气湍流短曝光降质图像的盲复原,本文在基于乘性迭代的盲图像复原算法的基础上进一步提出了一系列新的盲图像复原算法。针对噪声降质图像的盲复原,提出了带低通滤波的乘性迭代盲图像复原算法。为改善算法收敛性,提出了基于非对称乘性迭代的盲图像复原算法。为避免无意义歧义解,提出了带惩罚项的非对称乘性迭代盲图像复原算法,以及将算法向多帧序列降质图像的盲复原推广而提出基于乘性迭代的序列图像盲复原算法。 本文通过大量的实验仿真验证了所提出的新的盲图像复原算法的有效性,证实了算法对大气湍流降质图像质量的改善。
英文摘要: Owing to atmosphere turbulence, there are many degradation phenomena to the image of electro-optical tracking system such as shifting, glinting and blurring etc, which bring on great difficulty to the object detecting, recognizing and tracking. Especially, as the requirements of detecting capability, positioning accuracy, and tracking accuracy of electro-optical tracking system are more and more high in many fields, the effect of atmosphere turbulence on image quality of electro-optical tracking system becomes more notorious, and overcoming the effect becomes the imperiously key problem of these systems. The paper is spread out focus on the theme of overcoming the effect of atmosphere turbulence, and carries on the research of blind image restoration of images degraded by atmosphere turbulence for the purpose of improving the image quality and performance of system by applying blind image restoration technology. The main research in the paper is the related theories of blind image restoration and new algorithms for blind image restoration. Blind image restoration means to seek the best estimation directly from the observed degraded image, with incomplete or even little knowledge of object image and point spread function of system (called as the function of degraded process, too). The goal of blind image restoration is to reduce or eliminate the degradation of observed image, so that image quality is improved. Because it does not require knowledge of point spread function, blind image restoration fits to overcome the effect of atmosphere turbulence. However, there is so little useful knowledge that blind image restoration is very difficult with resolving. Considering the trait that the point spread function of many imaging system is the function of G class (including the point spread function of atmosphere turbulence long time exposal image), the paper proposes the DAPEX based blind image restoration algorithm for the blind image restoration of the long time exposal image degraded by atmosphere turbluce or other G class degradation process. The algorithm can estimate the point spread function directly from the degraded image by introducing the model of image power spectra. For the problem that noise is amplified in the restoration, the blind image restoration algorithm based on pre-denoising and DAPEX is proposed to achieve image restoration and restain of noise amplification simultaniously. Actually, the instantaneous movement of atmosphere turbulence is so stochastic and complex that it is hardly to be described. Therefore, the blind image restoration of short time exposal image degraded by atmosphere turbulence is even more difficult. To deal with the problem, the multiplicative iterative based blind image restoration algorithm is proposed in the paper. Owing to the product form of alternative estimate, the algorithm naturally preserves the nonnegativity of solutions if only the initial estimates of solution are nonnegative. This property of the algorithm makes it effectively avoid the computing complication of introducing penalized term for nonnegativity or imposing nonnegativity to solutions which usually occurs in other blind image restoration algorithms. In addition, unlike the classical image restoration algorithms which have division operation, such as RLA (Richardson-Lucy Algorithm) and ISRA (Image Space Reconstruction Algorithm) and so on, the algorithm is free from computing instability. It should be pointed out that the algorithm can achieve high resolution of image restored from single frame degraded image. In fact, the algorithm provides a new scheme for blind image restoration. For the blind image restoration of short time exposal image degraded by atmosphere turbulence, a series of new algorithms are derived from the multiplicative iterative based blind image restoration algorithm. For noisy degraded image, multiplicative iterative algorithm with low-pass-filter is proposed. To improve the convergence of the algorithm, the asymmetric multiplicative iterative based blind image restoration algorithm is presented. To avoid meanless divergent solution, the penalized asymmetric multiplicative iterative based blind image restoration algorithm is offered. Improving the algorithm and applying it to blind restoration of multiframe images, the multiframe multiplicative iterative algorithm for blind image restoration is provided. Through many exeperiments in the paper, the effectivity of these new proposed algorithms for blind image restoration is demonstrated and the improving of the quality of the image restored by these new algorithms is proved.
语种: 中文
内容类型: 学位论文
URI标识: http://ir.ioe.ac.cn/handle/181551/288
Appears in Collections:光电技术研究所博硕士论文_学位论文

Files in This Item:
File Name/ File Size Content Type Version Access License
10001_200518015127025张建林_paper.pdf(2829KB)----限制开放View 联系获取全文

Recommended Citation:
张建林. 成像跟踪系统中的盲图像复原研究[D]. 光电技术研究所. 中国科学院光电技术研究所. 2008.
Service
Recommend this item
Sava as my favorate item
Show this item's statistics
Export Endnote File
Google Scholar
Similar articles in Google Scholar
[张建林]'s Articles
CSDL cross search
Similar articles in CSDL Cross Search
[张建林]‘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
文件名: 10001_200518015127025张建林_paper.pdf
格式: Adobe PDF
此文件暂不支持浏览
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  中国科学院光电技术研究所 - Feedback
Powered by CSpace