Researches On The Methods Of Image Denoising Based On High-order PDE

In recent years,based on partial differential equations(PDEs) has been the most widely used method for image processing.There are rich theoretical and numerical methods for the partial differential equations,all these provide favorable conditions for the application of PDE.Because of that PDE is an adaptive denoising method,so it can choose the smooth degree according to the different areas feature of the image, it can effectly eliminate the noise in image,and at the same time preserve the image edge details.This paper introduces the ralative knowledge of image denoising methods based on PDE and introduce some classical denoising models according to the time order.As to overcoming the drawback called the “stair-case” that produced when use the second-order models to denoising,then put forward several higher order models for a particular noise based on existing fourth-order denoise models,the main innovation works are as follows:(1)According to the distribution characteristics of poisson noise based on fourth order Y-K model,then construct a new fourth order model.As solve and simulation experiment,the result shows that the new model gets the good effect at denoising work.(2)To overcome the drawback of “stair-case”,according to the characteristics of the speckle noise distribution,and put forward a kind of fourth-order denoising model.After solving the model and simulation,the experiment shows a good result.At the same time,the experiment on the medical B ultrasonic image which contains the speckle niose also gets a great effect.(3)The paper uses an efficient parallel denoising algorithm to deal with the new models which proposed before.The method using the ideas of parallel need to deal with the task into several parts,and send them to each computer,then integrate the independent results together to lead out the final solution.And adopt the parallel working on image denosing will greatly improve the efficiency of the computer.