摘要去马赛克算法是当今数码相机产品中的核心技术,算法性能的优劣直接影响图像的质量。因此,研究有效的去马赛克算法具有重要的应用价值。
本文是一篇对彩色图像去马赛克算法进行综述的文章,文章中分析比较了各种去马赛克算法的优点与不足。论文首先介绍了彩色图像处理的相关基本知识,然后主要介绍几种经典的以及最近比较流行的去马赛克的方法。主要有双线性插值方法,基于一阶和二阶微分的方法,色比恒定的方法,块匹配方法,非局部自相似驱动的方法等。在此基础上,分别对块匹配算法以及自相似驱动去马赛克算法进行改进。提出了两种改进的图像去马赛克算法。为了评估算法的性能,本文对多幅测试图像进行了插值处理,分别从主观视觉效果和客观图像质量评价指标两方面对算法进行了分析比较。实验结果证明了所提出的改进的非局部自相似方法具有比较好的去马赛克性能。68014
毕业论文关键词:Bayer滤波阵列 去马赛克 非局部
Abstract
Demosaicking algorithms are one of the most important technologies in digital camera products nowadays.The performance of algorithms directly affect quality of images . Therefore, researching on effective demosaicing algorithm has important application value.
This is an article about summarizing the demosaicing algorithms of color image , and we analysis advantages and disadvantages by comparing various algorithms in the paper. The paper firstly introduces the basic knowledge of color image processing, And then the paper introduces several classical and currently popular methods. Mainly algorithms have the bilinear interpolation method, based on first order, and two order differential, the block matching method, non-local self-similar driving methods. On this basis,I improve block matching algorithm and self-similar driving demosaicing algorithm. Put forward two kinds of improved image demosaicing algorithms. To evaluate the performance of the algorithm, this paper carried out interpolation processing on multiple images,from the subjective visual effect and objective image quality evaluation index of the two algorithms are compared and analyzed. The experiment results show that the improved non-local self-similarity method is a good demosaicing algorithm.
Keywords: Bayer CFA,DEMOSAICING,non-local
目录
第一章 绪论 5
1.1研究的背景以及意义 5
1.2去马赛克算法研究现状 7
1.3本文的工作安排 8
第二章 图像的预备知识 9
2.1 彩色模型 9
2.2 Bayer图像的视觉机理 12
2.3 Bayer模板CFA图像模型 13
2.4 常见的插值失真 15
第三章 CFA插值算法介绍 16
3.1 图像的相关性 16
3.2 双线性插值 17
3.3色比恒定 18
3.4 一阶微分边缘插值 19
3.5 二阶微分插值算法 20
3.6 块匹配算法 22
3.7 自相似应用颜色去马赛克 23
第四章 改进算法 26
4.1 改进的块匹配算法 26
4.2 改进的Self-similarity Driven算法 彩色图像去马赛克算法综述:http://www.youerw.com/shuxue/lunwen_76405.html