电子信息工程专业毕业论文-数字图像修复算法的研究

数字图像修复算法的研究

专业：电子信息工程

班级：

姓名：

目录

引言 3

1数字图像修复的概况 ............................................................................................ 6

1.1图像修复的研究背景与意义 ............................................................................................... 6

1.2图像修复算法的国内外研究现状 ....................................................................................... 8

2图像修复理论 ...................................................................................................... 12

2.1认知心理学格式塔理论与图像修复 ................................................................................. 12

2.2图像修复的三大原则 ......................................................................................................... 14

2.3图像修复效果的评判标准 ................................................................................................. 15

3 经典算法介绍 ..................................................................................................... 17

3.1基于SOM网络的图像修复 ............................................................................................... 17

3.2基于纹理合成的图像修复 ................................................................................................. 22

4基于TV模型的数字图像修复 ........................................................................... 31

4.1TV修复模型的介绍 ............................................................................................................ 31

4.2 TV修复模型的基本原理 ................................................................................................... 33

4.3 TV修复模型算法的数值实现 ........................................................................................... 35

4.4实验结果与分析 ................................................................................................................. 40

结论 ......................................................................................................................... 47

致 谢 ..................................................................................................................... 48

参考文献 ................................................................................................................. 50

附录1 源程序清单 ................................................................................................ 52

摘要

对图像的损坏或缺失区域进行修复在数字图像处理领域是一个重要课

题。同样在信号处理、计算机视觉和计算机图形学也是一个热点问题，主要任务是让计算机自动按照人眼视觉特点，根据图像破损区周边的信息来修复指定的破损区。

数字图像修复是对图像中信息缺损的区域进行的修复，其主要目标是修复有信息缺失的部分，并使人们没有办法发觉图像曾经破损或者已经被修复。在该领域的研究，国外的发展要比国内的发展好的非常多。为了提高受损数字图像的修复质量，将整体变分模型用在图像修复中。本文阐述了全变分修复模型的原理数的实现算法。基于Matlab编程软件对修复过程仿真的实现，克服了直接求出偏微分方程的难题。仿真实验结果证明，这种算法收敛快、修复图像的视觉质量也非常好，特别适合线状裂痕的图像修复。随着迭代次数的增长，被修补区域贴近原图的质量，当破损区域较小而且受损区域灰度梯度不大的时侯，不会出现明显的修复痕迹。

关键词：

破损图像；全变分；算法收敛；数字图像修复；偏微分方程

沈阳大学毕业设计（论文） NO 1

Abstract

Reconstruction of missing or damaged regions of images, known as inpainting or retouching is a very important topic in image processing and a hot spot in signal processing, computer vision and computer praphics. Its main work is to make a computer automatically inpaint appointed damaged region in a digital image according to the information of the regions around the damaged regions and characteristics of human vision.

Image restoration is to repair the damaged area of the image, its main task is to repair the part of information deletion, and make people have no way to find the damaged images have been repaired. In the field of research, the development of foreign is better than the development of the domestic. In order to improve the inpainting quality of the damaged image, the TV model is used for image restoration. TV model are introduced in this algorithm, and the application of model. The TV model theory, structure and numerical algorithm are introduced. The process of repairing is programmed based on matlab software, overcomes the problem of partial differential equations directly. Simulation experiment results show that this algorithm converges fast, the repair quality is very good, especially suitable for image restoration of the linear cracks. As the growth of the number of iterations, the repaired area is close to the original image quality, when the damaged area is small and the damaged area is small and the damage area gray gradient is not big, it cannot appear obvious trace.

Keywords：

damaged image；total variation；algorithm convergence；digital image inpainting； partial differential equation.

沈阳大学毕业设计（论文） NO 2