Research Of Image Denoising Based On The Multi-wavelet

In the process of acquisition and transmission, the images often get interference from various factors. So images denoising is an important study for image processing. In recent years, due to its orthogonality, symmetry, smoothness, compact support , the multi-wavelet following the wavelet was received wide attention and widely used in the signal areas including the image processing. Although there are many methods for the image denoising, it's just get concerned for the image denoising based on the multi-wavelet, which has great research significance both in theory and practice.In this paper, the author made a deep research about the image denoising based on the multi-wavelet. The following is included. Firstly, the author introduced the wavelet and multi-wavelet in details. Secondly, the traditional image denoising methods are introduced, particularly the threshold shrinkage denoising base on the wavelet and multi-wavelet. Furthermore, the author gave some algorithms in details. Thirdly, because of the deficiencies of the threshold shrinkage denoising method based on the wavelet, this paper gave analysis in detail about a improved Wiener filtering algorithm. Which used the relationships of the wavelet coefficients and the result of the additive Gaussian noise being Gaussian distribution at different scales and different directions. Lastly, the author improve the above method. Using the characteristics of the image wavelet coefficients being established, the decomposition of the wavelet coefficients being more detailed, and noise variance estimation being more accurate, it was proposed that images are transformed to multi-wavelet coefficients and are filtered by the improved wiener filter. The test with the software of Matlab showed the validity of the method. This paper mainly described the theories of the image denoising based on the multi-wavelet and some key technologies such as image preprocessing and multi-wavelet transforming. From which we can see its advantages. In this paper the author gave detailed description of the image denoising steps using GHM multi-wavelet. What's more, the CL multi-wavelet and multi-wavelet are introduced. Form all the above a good foundation was made for the future.As the Wiener filter is an adaptive filter, it's particularly good for the Gaussian noise to be denoised. It can just output the image according to the regional variance. The variance is greater, the image is more smooth. After the images are transformed to multi-wavelet coefficients and are filtered by the improved wiener filter, the better effects will be got.