Research On Image Deniosing Model Of Partial Differential Equations

In the modern society,science and technology to produce,the images help mankind to observe and feel objects in different environment s and scals,and to make optimal decisions. However, in reality the image sensor will be affected by various factors, resulting in a variety of random pulse interference and other noise, the image in the process of formation, transmission and communication, often accompanied by them, whitch affect the quality of the image and bring a lot of inconvenience of subsequent processing and analysis. Therefore, it is particularly important for image denoising in the image preprocessing.However, it is a pair of contradictory relationships on image smoothing and edge detail to keep. Traditional denoising methods, such as Gaussian filtering, mean filtering,they remove the image noise while blurring the edges of the image. Denoising method based on partial differential equations is an adaptive denoising technology, in the process of denoising, image smoothing and feature detection simultaneously, taking into account noise filtering and features remain,that is a good image denoising technology.In this paper, I research on the image denoising model under the theoretical framework of partial differential equations, by adopting the combination of theoretical analysis, simulation and experimental method, and discuss some difficulties in the model. Based work includes the following:1) Describes the image denoising based on partial differential equations (PDE) of mainstream models, including the vector diffusion model, the tenser diffusion model and the forward-backward diffusion model, and the advantages and disadvantages of these models are analyzed, Meanwhile the simulation experiments are carried out.2) Elaborated on the characteristics of the diffusion coefficient of the PM model.The disadvantages of the PM model are analyzed theoretically, and by improving the diffusion coefficient, make it can filter out noise while preserving image edges better detail features, to some extent to overcome the contradiction between the edge preserving and noise elimination, reducing the image " block ".3) Elaborated on the mechanism of the denoising of fourth-order model, for the phenomenon which can not remove salt and pepper noise,carry on the thorough analysis, and by improving the diffusion coefficient and improve the performance of the model.4) Construct a new model for the advantages and disadvantages of the second-order model and fourth-order model, which models the use of PM characteristics of fast iteration, and fourth-order model can eliminate the " blocking effect" caused by the second-order model characteristics, not only speed up the smooth speed, but also improves the quality of the image.