Research On Image Derioisine Based On Wavelet Transform

Human access to information by the human eye, the human eye can see visualinformation, the main image information, which means that the image is our mainsource of information. However, image capture and transfer process easily by outsidenoise pollution, the presence of noise so that the image can not effectively identifycertain characteristics details. So, in the neighborhood of the image processing, howcan effectively remove the noise is particularly important, denoising subsequentimage processing such as image compression, signal detection and feature extractionis a necessary step. Improve the image signal-to-noise ratio of the maximum to retainthe characteristics of the original image, the main goal of our study of image noisereduction.Wavelet theory is local time-frequency analysis, time domain and frequencydomain joint, the development is very rapid, the wavelet analysis is a powerful tool ofthe non-stationary signal processing signal by zooming and panning operationfunction, he carried out a detailed analysis of the multi-scale calculation can beeffectively extracted from the signal information. Wavelet analysis is internationallyrecognized as the high-tech fields, is a cutting-edge topics and research focus in thefield of signal and information processing, and therefore the image signal filteringnoise, to attract the more image edge detection, image compression, image fusionapplications to more researchers.In this paper, the study image denoising wavelet analysis theory, take advantageof the special advantages of wavelet analysis to process image information. Thisthesis work will commence in the following aspects: First, the traditional imagedenoising method described in detail, the history and development status of waveletimage denoising, wavelet denoising description of the problem and the advantagesand disadvantages, and how to evaluate a de-noising algorithm the pros and cons.Then introduced the basic theory of wavelet analysis, wavelet analysis of the course of development and the range of applications, introduces its most representativemulti-resolution analytical knowledge. Secondly, the details of the waveletcoefficients of the model of the three main applications of wavelet denoising process,scale model, scale model and the hybrid model, as well as the common types of noisein the image, and their respective characteristics. Then, based on wavelet analysis ofseveral most classic of threshold denoising method threshold value denoising,threshold denoising, threshold denoising and threshold value denoising, as well asanalysis of the main threshold denoising method most critical threshold value of thefunction and the threshold value selection problems, this paper presents two newnoise reduction algorithm on the basis of the classical threshold, a wavelet thresholddenoising algorithm, MATLAB simulation test, the denoising effect has improvedgreatly. Also proposed a simulation based on adaptive neighborhood coefficients ofwavelet image threshold denoising method, the second method the same gooddenoising effect and denoising is better than the first method.