Image Denoising Algorithm Based On Gaussian Noise

During acquisition or transmission,image will be corrupted inevitably by all kinds of noises,which are the killers of quality of images and obstacles to the later work, such as image segmentation, feature extraction,object recognition,etc.In order to alleviate or eliminate the effects of noises,many approaches are proposed by relevant talents. For example, recovering the original image from the noise-polluted image.In general, denoising poses a trade-off problem between noise removal and preservation of detail. Many a denoising algorithms are proposed up to now,but some of them are poor to remove the noise while retaining the important signal features and details or are more time-consuming.It has the practical significance to study suppressing the Gaussian noise for the reason that Gaussian noise is one of the principal noises.In order to reach a satisfactory trade-off between noise suppression and detail preserving, and have the superiority in computing speed, a novel denosing algorithm for a single image corrupted by Gaussian noise is proposed after some systematic investigation is given to the local variance of dirty image,advantages and disadvantages of mean filter and wiener filter and the capactiy of anti-noise of image segmentation.The algorithm has two merits,one is that it can preserve the image edge well while denoising,the other is that it has a faster speed and is adaptive to different noise environments.The experimental results demonstrate that the algorithm not only has the better capable of suppressing noise preserving the details than the tranditonal algorithms,but also has the faster speed than some algorithms.