【正文】
方面達(dá)到了明顯的消噪效果,有很好的應(yīng)用前景。圖像融合在信息融合中是重要的,在目標(biāo)識別、機(jī)器視覺、智能系統(tǒng)、醫(yī)學(xué)圖像處理等領(lǐng)域被廣泛應(yīng)用。基于小波變換的多分辨率分析算法則是在頻率域?qū)崿F(xiàn)了圖像的融合,有效幫助理解圖像并快速獲取感興趣的信息。通常有兩種融合方法:簡單融合法和參數(shù)獨(dú)立法。 實(shí)現(xiàn)融合的主要函數(shù)在MATLAB中實(shí)現(xiàn)圖像融合的函數(shù)是wfusimg,其調(diào)用格式為:XFUS=wfusimg(X1,X2,WNAME,LEVEL,AFUSMETH,DFUSMETH)說明:返回將原始圖像X1和X2融合后的圖像XFUS,參數(shù)LEVEL是指X1和X2分解的層次,參數(shù)WNAME指定分解小波,矩陣X1和X2的大小必須相同。使用sym4小波對待融合圖像進(jìn)行5層小波分解,獲得得相應(yīng)的分解系數(shù),并取細(xì)節(jié)和近似信號相應(yīng)系數(shù)的最大值利用融合函數(shù)wfusimg進(jìn)行融合,最后重構(gòu)并顯示融合后的圖像。除此之外,通過小波變換也可以實(shí)現(xiàn)兩幅模糊圖像的融合。基于小波變換的圖像融合可以應(yīng)用在采用不同成像機(jī)理得到的同一物體部件的圖像上,例如:多頻譜圖像理解、醫(yī)學(xué)圖像處理等。簡單介紹了從傳統(tǒng)傅立葉變換到小波變換的技術(shù)發(fā)展,體現(xiàn)小波變換在圖像處理上的優(yōu)越性。簡單扼要地介紹了一些處理圖像的關(guān)鍵小波函數(shù)的調(diào)用方法,體現(xiàn)運(yùn)用小波變換對算法的簡化效果十分明顯。本文算法相對較為簡單明了,雖然有待進(jìn)一步對小波變換理論深入研究,但卻已然表現(xiàn)了小波變換和傳統(tǒng)變換相比的優(yōu)越性,同時(shí)體現(xiàn)了小波變換已經(jīng)可以廣泛應(yīng)用在圖像處理領(lǐng)域中,并占據(jù)重要作用,擁有廣大的發(fā)展前景。在本文完成之際,衷心感謝我的指導(dǎo)老師楊藝敏老師的教導(dǎo)和幫助。最后,感謝我的母校的傳道、授業(yè)、解惑。感謝所有在大學(xué)四年互幫互助的同學(xué),讓我在陽光和笑容中收獲珍貴的友誼和寶貴的知識。 or the way a digital camera processes visual scales of resolutions, and intermediate details. But the same principle also captures cell phone signals, and even digitized color images used in are of real use in these areas, for example in approximating data with sharp discontinuities such as choppy signals, or pictures with lots of edges. While wavelets is perhaps a chapter in function theory, we show that the algorithms that result are key to the processing of numbers, or more precisely of digitized information, signals, time series, movies, color images, etc. Thus, applications of the wavelet idea include big parts of signal and image processing, data pression, fingerprint encoding, and many other fields of science and engineering. This thesis focuses on the processing of color images with the use of custom designed wavelet algorithms, and mathematical threshold filters. Although there have been a number of recent papers on the operator theory of wavelets, there is a need for a tutorial which explains some applied tends from scratch to operator theorists. Wavelets as a subject is highly interdisciplinary and it draws in crucial ways on ideas from the outside world. We aim to outline various connections between Hilbert space geometry and image processing. Thus, we hope to help students and researchers from one area understand what is going on in the other. One difficulty with municating across areas is a vast difference in lingo,jargon, and mathematical terminology. With handson experiments, our paper is meant to help create a better understanding of links between the two sides, math and images. It is a delicate balance deciding what to include. In choosing, we had in mind students in operator theory,stressing explanations that are not easy to find in the journal literature. Our paper results extend what was previously known, and we hope yields new insight into scaling and of representation of color images。 here, Wavelet Toolbox is used [Gon04]. Digital Image Representation. An image is defined as a twodimensional function ie. a matrix, f(x, y), where x and y are spatial coordinates, and the amplitude of f at any pair of coordinates (x, y) is called the intensity or gray level of the image at the point. Color images are formed by bining the individual twodimensional images. For example, in the RGB color system, a color images consists of three namely, red, green and blue individual ponent images. Thus many of the techniques developed for monochrome images can be extended to color images by processing the three ponent images individually. When x, y and the amplitude values of f are all finite, discrete quantities, the image is called a digital image. The field of digital image processing refers to processing digital images by means of a digital puter. A digital image is posed of a finite number of elements, each of which has a particular location and value. These elements are referred to as picture elements, image elements, pels and pixels. Since pixel is the most widely used term, the elements will be denoted as pixels from now on. An image maybe continuous with respect to the x and ycoordinates, and also in amplitude. Digitizing the coordinates as well as the amplitude will take into effect the conversion of such an image to digital form. Here, the digitization of the coordinate values are called sampling。 reads the JPEG image lena into image array or image matrix f. Since there are three color ponents in the image, namely red, green and blue ponents, the image is broken down into the three distinct color matrices fR fG and fB。 that agrees with what was discussed above in Daubechies wavelets that db2 is being better in signal pression than db1(Haar). Considering the errors and pression ratios as well as the perception of the image sym5 would be the best choice of wavelets, among the ones that was used for the image pression. So, in this case, sym5 being very close to symmetric wavelet did a better job in image pression. Also, having the extra properties as mentioned under the Coiflets section made Coif3 perform better in image pression. Having biorothogonal property also seem to result in better image pression. On the other hand the orthogonal Daubechies wavelets do not seem to perform better than coiflets, biorthogonal wavelets and , having longer support which is proportional to the order of the wavelet,appears to worsen the performance of the image the threshold value 10, when a Daubechies wavelet, db1 was used the pression ratio was while db2 resulted in . A Coiflet Coif1 resulted in pression ratio of whereas Coif 3 resulted in pression ratio of . Biorthogonal wavelets and gave and for the pression ratio respectively. Symlets sym2 and sym5 resulted in pression ratios of and respectively. Now, with higher threshold value, sincemore date is being lost, the pression ratio increases. However, the quality of the image diminishes at the same wavelet deposition of images was performed the number of times the image can be divided by 2 ie. (floor(log2(min(size of Image)))) times. The averaged image of the previous level is deposed into the four subimages in each level of wavelet