Image Denoising Method Based On Wavelet Transform

Images are disturbed inevitably by all kinds of noise when collected and transmitted. In order to higher level processing, it's very necessary to carry on de-noising pretreatment to images. In recent years, the wavelet theory has obtained the extremely rapid development. For the characteristics of its low entropy, multi-analysis, relativity removal and flexible bases, the wavelet transform has become a powerful tool in the field of image de-noising.The development of wavelet theory and its application situation in image processing are introduced. Then, classical wavelet image de-noising methods which are often used at present are described in details. These algorithms are compared and the principle, characteristic and disadvantage of each algorithm are also analyzed.Firstly, as to Gaussian noise ubiqtious in image, an improved threshold based on region correlation is formed. This method reserved some details which are overskilled in Visushrink by using wavelet region correlation. It has overcome'overskilled' disadvantagement. Then, to images containing Gaussian and impulse noise, if it is directly processed in wavelet domain, the threshold will be too high, and this will result in 'overskilled'. Concerning this problem, we use extreme median filter to remove impulse noise. Then in wavelet domain, we use wavelets in each region to compute local threshold. The new threshold is better in local region and it has overcome the 'overskilled' shortcoming.The results of the simulation experiments show that the new algorithms are superior to the classical threshold image de-noising methods. The new method can not only effectively remove noise but also preserve edge details information, so it is efficient in image de-noising.