文档介绍：第一代曲波的图像处理分析
基于第一代曲波变换的图像处理分析
摘要
最近,由于小波变换具备良好的空间域与频域的局限性,所以在图像去噪方面得到了广泛的应用。但是,小波变换的特点只是是反映信号的点奇异性,而不能最优表示含线奇异或者面奇异的高维函数。因此,需要有比小波变换更优化的函数表示方法,以充分利用图像的几何特征。
曲波变换是一种可以很好的逼近含线奇异的高维函数,所以最近几年受到广大研究人员的普遍关注和重视。但是传统曲波变换往往是冗余度较高、非正交性的,并且曲波变换算法复杂,在变换过程中速度较慢。针对这一特性,本文利用的曲波变换对物体边缘信息具有最优稀疏表示。通过大量实验表明,基于曲波变换的图像消噪算法可以很好的去除绝大多数类型的噪声,能够较好保留图像边缘信息,使去噪后的图像达到很好的效果,同时在图像压缩、融合等图像处理方面也变现了较好的处理效果。
关键词:小波变换;曲波变换;
Abstract
Recently, as a result of wavelet transform good space domain and frequency domain, so the limitations of the image denoising has been widely used. However, the characteristics of wavelet transform is reflected signal only point singularities, not optimal says contain line singular or face singular large-dimensional function. Therefore, we need to have more than the wavelet transform the function optimization method, so as to make full use of said the geometrical characteristic of the image.
Curvelet transform is a kind of very good approximation of high dimensional function containing line singular, so in recent years by the researchers the universal attention and attention. But the traditional curvelet transform is often redundancy taller, the orthogonality and curvelet transform algorithm in the transformation plex, the more slowly. According to its characteristic, this paper using fast curvelet transform (FDCT) overcame curvelet transform algorithm of object with optimal edge information sparse said. Through a large number of experiments show that, based on the image denoising algorithm FDCT can be very good remove most types of noise, better able to retain the image edge, Make the image denoising to good effect.
Key words:Noise;Wavelet transform ; Curvelet transform ;
目录
摘要 I
Abstract II
前言 1
第一章绪论 2
研究背景和意义 2
国内外研究现状 3
第二章小波去噪的基本原理和应用 4
小波去噪中的小波系数模型 4
层间模型 4
层间模型 4
混合模型 5
各种小波变换在小波去噪中的应用 5
第三章曲波变换变换理论 7
(Curvelet)变换 7
第四章 8
第五章算法实现 9
结论 10
参考文献 10
致谢 12
前言
近年来,小波变换在图像处理中得到了成功的应用,但小波变