The Remote Sensing Image Processing Based On Partial Differential Equation

The remote sensing image noise reduction technology is the key problem in the remote sensing image processing.In a degree noise can decline the clarity and the quality of the image. Noise reduction technology can improve the quality of image, make analysis and recognize the image easily, and make remote sensing image can get more extensive application.The method based on partial differential equations(PDE) in image processing is one class of image processing methods, which is developing rapidly in recent years. It is a new mathematical tool used in image processing after the wavelet transform image processing method. It becomes an important method in the field of image processing. The PDE image processing method is based on mathematical theory and can unify with traditional image processing method, so development and application prospects are all very extensive.Firstly, this paper analysis remote sensing image and its noise simply, and then introduces the theoretical basis of the PDE and the typical partial differential equation methods that are used in image processing. It is easy to produce many regions of constants and false edge in image processing results. This makes the visual effect very unnatural. This phenomenon is named "ladder effect". This paper focuses on researching the typical method to eliminate the "ladder effect" with the four-order PDE method, and through the experiments to analysis and test the method. For high order PDE to eliminate the "ladder effect " has shortcomings such as:The results’ space of this kind of equation do not belong to the BV function space, this leads to the results to be excessive smoothing easily, so as to produce boundary leakage. This paper also puts forward the models that based on gradient of fidelity denoising. And it uses the tested images to simulate and validate the algorithm, compares the denoising results with other image denoising algorithm. Through the numerical results evaluate the denoising effectiveness, the conclusion is drawn that the method of this paper can not only remove the noise and be able to retain the image edge information better.This method makes up for the defect of higher order nonlinear diffusion denoising method, which is fuzzy boundaries, and improves the quality of the remote sensing image, ameliorates the visual effect.