Research And Realization Of Image Denoising Based On Wavelet Transform

The image denoising is an eternal theme of image processing research. The image is often corrupted by noise in its acquisition or transmission. The noise to be removed before analysis has an important effect on image analysis. Image-denoising is an important technology in image analysis and processing domain.Wavelet thresholding denoising method is one of the best denoising methods of image, which uses the wavelet decomposition, from the different features of sub-band images, selecting a different threshold, so as to achieve better denoising results. Wavelet transform is the following tools of time-frequency analysis after Fourier transform. However, Wavelet transform in the time-frequency domain has a good nature and the localization of the characteristics of multi-resolution analysis, and,it can not only meet the requirements of the various de-noising, such as low-pass, high pass, the removal of random noise, and compared with the traditional denoising methods, has unparalleled advantages, it become a powerful signal analysis tool, known as the mathematical analysis of signals microscope. Applications include image preprocessing, image compression and transmission, image analysis, feature extraction, such as a lot of image processing stage, the removal of Gaussian noise has achieved very good results.Wavelet transform has a "mathematical microscope" role in the denoising while preserving image details which can help to restore the original image. In a large number of wavelet denoising methods, the most use is Donoho wavelet shrinkage threshold method, but the Donoho threshold has larger reconstruction error.Based on this idea, this paper introduce a image denoising method based on the combination of wavelet transform and median filtering. Before denoising, wavelet edge detection determine the edge of the characteristics of image wavelet, preserving the wavelet coefficients in these locations, avoiding the impact of other adaptive wavelet coefficient threshold denoising,such as Gaussian noise. And,then using the median filter to remove salt and pepper noise. In the practical computer programming,it also involves the wavelet function design of the data structure, the choice of wavelet bases, as well as variety processing of wavelet. In the fifth chapter will introduce the wavelet data structure, the definition of wavelet bases, and how to realize the wavelet transform.The experimental results show that the algorithm can not only filter out image noise mixture of Gaussian noise and salt and pepper noise, but also reserve the details of the edge of image. The filter effect is better than the traditional method of image denoising. And introduce the computer simulation of image denoising based on wavelet transform by software MALTLAB.