The Mathematical Modeling Wavelet And PDE On SAR Image Denoising

Since Synthetic Aperture Radar (Synthetic Aperture Radar, SAR) have the characteristics of all-weather, all-day when imaging, high resolution and the strong penetrability, now has been widely usedin the aspect of national defense construction and economic construction. But the presence of specklenoise of SAR image is seriously hinder the subsequent image processing, and it can hinder extract thefeatures in the image data information, so how to effectively to filter speckle noise is a very key stepsof the SAR image preprocessing work. It has very important theoretical value and practical applicationvalue.The difficult problem of SAR image speckle noise reduction is how can get best denoising effect,and can keep the image details such as the edge and texture information. SAR speckle noise filteringmethods are divided into two categories: first, visual processing filter; Second, filtering processingmethod after imaging, and can be divided into three categories, the denoising methods based on spacedomain, the denoising method based on wavelet domain and denoising method based on partialdifferential equation.This article focuses on the SAR imaging denoising method, then analyzes the SAR speckle noisebased on wavelet domain transform filtering method, and discusses the advantages and disadvantagesof wiener filtering, according to the needs of this paper to establish mathematical model. This paperproposes a new wavelet domain wiener filtering algorithm, and simulation results verify theeffectiveness of the proposed algorithm.Then introducing the SAR speckle noise filter based on partial differential equation method, theanalysis summarized a variety of classical partial differential equation of denoising method which wasput forward by the predecessors on the basis of a composite model of partial differential equations ispresented. This model is to use two weighting function to P-M equations, direction of diffusionequations and LLT combine the high order partial differential equation model. This model is to keepthe second order and fourth order partial differential denoising model, the advantages of slam the doormodel disadvantages to effectively improve the ability to keep the image edge details and obtain gooddenoising effect.Finally, using mathematical modeling methods to improve the new wavelet domain wienerfiltering and partial differential equation, a comprehensive model combined with MATLAB softwarefor SAR image processing. The experimental results show that the mathematical model can get rid ofdisadvantages of wavelet transformand the partial differential equation, not only can filter out most ofthe speckle noise, but also can effectively keep the detail of the image characteristics, and to improvethe operation efficiency.