【正文】
信息的媒介,對(duì)圖像信息的認(rèn)識(shí)理解是由人的視覺系統(tǒng)所決定的。 目前大多數(shù)數(shù)字 圖像系統(tǒng)中,輸入光圖像都是采用先凍結(jié)再掃描方式將多維圖像變成一維 電信號(hào),在對(duì)其進(jìn)行處理、存儲(chǔ)、傳輸?shù)燃庸ぷ儞Q,最后往往還要再組成多維圖像信號(hào)。 但在很 多情況下,這樣描述方法是很復(fù)雜,甚至不可能的,而實(shí)際應(yīng)用往往也不必要,通常使用其數(shù)值特征,即均值方差、相關(guān)函數(shù)等。 對(duì)灰度圖像 ? ?,f xy 來(lái)說(shuō) , 可看做是二維亮度分布,則噪聲可看做是對(duì)亮度的干擾,用 n( x, y)來(lái)表示。一幅圖象在實(shí)際應(yīng)用中可能存在各種各樣的噪聲 ,這些噪聲可能在傳輸中產(chǎn)生 ,也可能在量化等處理中產(chǎn)生 。實(shí)驗(yàn)結(jié)果表明,基于脊波變換的圖像去噪和融合方法具有比小波變換更好的效果。其次,針對(duì)圖像去噪中常用閾值方法的缺點(diǎn)和不足,提出了一種基于脊波變換的改進(jìn)的圖像去噪算法,該算法采用指數(shù)型閾值函數(shù),利用 sureshrink 自適應(yīng)閾值。分別論述了小波變換和脊波變換基本理 論,基于脊波變換的圖像去噪以及圖像融合。 畢 業(yè) 設(shè) 計(jì) 題 目: 基于脊波變換的圖像去噪研究 所在 專業(yè) : 電子信息科學(xué)與技術(shù) 學(xué)生簽字 : _______________ 導(dǎo) 師 簽字 : _______________ II 摘 要 圖像中的噪聲影響圖像的輸入、采集、處理的各個(gè)環(huán)節(jié)以及輸出的全過程,尤其是圖像輸入、采集中的噪聲必然影響處理全過程以至最終結(jié)果,因此在圖像預(yù)處理中必須減少圖像中的噪聲。 本文以脊波變換為研究對(duì)象,論述了脊波變換在圖像處理中的應(yīng)用。首先,在分析小波變換理論的基礎(chǔ)上,結(jié)合小波變換的優(yōu)缺點(diǎn),為了克服小波變換在圖像處理中的不足,介紹了脊波變換的基本理論。最后,將脊波變換的思想應(yīng)用于圖像融合,采用區(qū)域方差的融合規(guī)則,得到了一種基于有限脊波變換的圖像融合算法。 關(guān)鍵詞 :脊波變換 小波 變換 圖像去噪 圖像融合 III ABSTRACT The image of the noise impact of the input, collecting, processing, output of the whole process, especially the image of the input, sources of noise is dealt with and influence the whole process and ultimate in image preprocessing, so we must reduce the noises in the image This paper deals with Ridelet Transform in processing, which involves the basic theory of Wavelet Transform and Ridelet Transform, finite Ridelet transform in image denoising and in image fusion. Firstly, depending on transform, for basic theory of Ridelet transform. Secondly, a improvement of image denoising algorithm based Ridelet transform is presented to overe the disadvantage and deficiency of the mon threshold method at image denoising. The exponential threshold function and the adaptive SureShrink threshold value are applied into this approach. Thirdly, Ridelet transform is applied in image fusion, adopted the fusion rule of regional variance, an image fusion algorithm based on finite Ridelet transform has appeared. The results of experiment indicate based on Ridelet gain better effects than wavelet transform. Key words: Ridelet transform wavelet transform Image Denoising Image Fusion 1 目 錄 摘 要 ................................................................................................................................. I ABSTRACT ............................................................................................................................. III 第 1 章 緒 論 ................................................................................................................. 3 圖像中的噪聲及去噪方法概述 ............................................................................ 3 圖像中的噪聲 ............................................................................................. 3 圖像去噪方法概述 ..................................................................................... 4 小波的發(fā)展 現(xiàn)狀及應(yīng)用前景 ................................................................................ 4 脊波的發(fā)展現(xiàn)狀及應(yīng)用前景 ................................................................................ 5 論文的研究?jī)?nèi)容與組織結(jié)構(gòu) ................................................................................ 6 第 2 章 圖像去噪及其發(fā)展 ................................................................................................. 7 傳統(tǒng)去噪方法 ........................................................................................................ 7 小波變換圖像去噪方法 ........................................................................................ 7 小波去噪發(fā)展歷程 ..................................................................................... 7 小波去噪方法 ............................................................................................... 8 本章小結(jié) ................................................................................................................ 10 第 3 章 小波分析基本理論 ............................................................................................... 11 小波變換基本理論 .............................................................................................. 11 連續(xù)小波變換 ........................................................................................... 11 離散小波變換 ........................................................................................... 12 二進(jìn)小波變換 ........................................................................................... 12 多分辨分析 ............................................................................................... 12 Mallat 算法 .............................................................................................. 13 圖像的小波 變換 ....................................................................................... 14 本章小結(jié) .............................................................................................................. 18 第 4 章 脊波變換 ............................................................................................................... 19 脊波變換基本理論 .............................................................................................. 19 連續(xù)脊波變換 ........................................................................................... 19 離散脊波變換 ........................................................................................... 20 脊波變換的實(shí)現(xiàn) .................................................................................................... 20 Ridgelet 變換與 Wavelet 變換的聯(lián)系 ............................................................. 21 有限 Radon 變換 .................................................................................................. 23 數(shù)字脊波變換 ...................................................................................................... 25 脊波變換的數(shù)字實(shí)現(xiàn) .....................................................................