Image Denoising Methods Based On Non-local Means And Non-local TV

In the process of acquisition and transmission,Image will be inevitably influenced by the noise.And it makes the image distortion,fuzzy,and the visibility of the image is affected.However,we usually require more accurate image information.At the same time,image denoising is the basis of many application areas,such as pattern recognition,digital entertainment,remote sensing imaging and so on,has the very important research value.A good denoising algorithm becomes a key step to improve the quality of final visual image.Therefore,scholars have been doing the related research.In recent years,a large number of new denoising algorithm is proposed.And the Non-local Means algorithm is one of the most popular algorithm.It was proposed by Buades in 2005.And since then,more and more scholars began to study of the algorithm based on the Non-local mind.It has a good denoising performance,but there are some disadvantages in the edge and texture protection for the algorithm.Based on this,the work of this paper includes two parts as follows:Firstly,this paper presents an image denoising method based on Bilateral Filtering and Non-local Means.Because image information is continuous,the image block which has a more closer distance with the target image block will contain more similar information.And it need to be given a bigger weight.Accordingly,this paper improves the weight of the Non-local Means method.This paper combines the geometric closeness factor with the weight of the Non-local Means,and constructs a new weight.It eliminate the influence of dissimilar pixels further.Then the structure of image has been better protected.Secondly,this paper presents an image denoising method based on Non-local TV model and gradient histogram preserve.This paper combines the gradient histogram preserve with the Non-local TV model,and presents an image denoising method based on Non-local TV model and gradient histogram preserve.Because histogram specification can protect texture features well.We hope that the new algorithm can have a good denoising performance.And at the same time,it can protect image texture details very well.