The Research Of Image Denoising Method Via PDEs

Digital image processing–the manipulation of images by computer– isa relatively recent development it terms of human's ancient fascination withvisual stimuli.In its shot history, it has been applied to practically every typeof imagery, with varying degrees of success. The inherent subjective appeal ofpictorial displays attracts perhaps a disproportionate amount of attention fromscientist and lay person alike.In the generation and transmission process of Digital Image , the variousnoise from the environment of equipment and a variety of impacts is a majorreason of the decrease of image quality, at the time this impact on digital imagecompression and transmission. Noise sources including electric sensor noise,photos particles noise, electromagnetic interference, channel error and so on.Therefore, how to get rid of the images noise has become an important task indigital image processing .In the traditional method of reducing noise, made image of the borderfuzzy, and the re?ect of people to high-frequency components (edge details)is very sensitive, the most of the information of image also come from some ofthe edge and contour. Therefore, finding a new effective algorithm has been ac-tive in this field research. The image processing methods based on PDEs in thisarea have been widely attention because of they can smooth noise and maintainthe edge.In allusion to the PDE method of Image Denoising, This article describes avariety of equation of de-noising, summarize PDE for the application of the gen-eral denoising principles and steps, and under the image's character coordinates(η,ξ), analyze the various models and summarized the nature of the essence ofvarious types of equations: In the edge of the image part. Equation is small along the directionη,ξof the diffusion coefficient, we can get the purposes of maintaining the imageedge information through control Diffusion coefficient by its law line.In image's smooth region, various equations accelerate smooth imagesby increasing its diffuse speed .About these lection of diffusion ,we coefficient general used decreasingfunction of the gradient. While the Improvement of coefficient should be re-ected in their sensitivity to noise, such as method catte and lin is the originalimage and gauss function do convolution to further reduce the noise on the de-gree of sensitivity of the coefficient.Ahout above analysis ,the authors propose ,we bring forward the followingmodel for Image Denoising:αenhanced diffusion coefficient is constants that is greater than 1 , k is canny noise operator, he is coefficient on the adaptive image .In the above equation along the edge of to the direction the image is basicnon-proliferation, along the tangential direction it is expansion this will keep theborder at the same time accelerate the proliferation of speed. Traffic smooth inthe region are expanding in two directions. In the cusp of the image we add anitem to protect it , it is effective to avoid the details of the loss of image pointed.Finally the author do a numerical the experiment about this model, resultsshow that the means to enhance the image SNR and achieve the anticipatedresults.