Fuzzy Image Denoising Based On Sparse Prior

High quality digital images is an essential prerequisite for the depth of the image application,but should be applied in the image and the image in the process of signal transmission,due to the defects of imaging equipment,low atmospheric visibility,relative motion between the target object and imaging equipment,the photographer jitter operation,transmission medium is not perfect and other factors,finally presents a digital image is pixels fuzzy,quality degradation,loss of information,the application of the fuzzy digital image of the quality problems in the field of image recognition and deep learning,the noise of the image pixels with mutual interference between pixels,correct application will lead to low efficiency,And then bring unpredictable economic losses and major disasters.So how to manage the depth fusion,image study and mathematical optimization recovery image denoising algorithms,effectively reduce the noise interference with the original image,improve the quality of blurred image,not only to promote the development of upgrade of fuzzy image denoising recovery technology with the theory and practice value,can also promote different industries,different disciplines,different areas of their progress and blend each other.At first,this paper involves in the fuzzy image denoising and restoration of related theory and common methods in detail,including the mechanism of image degradation model,denoising model,the cause of the blurred image with prior knowledge of classification,the image denoising recovery technology,for the subsequent algorithm design and research of theory and method of foundation.Secondly,based on the internal mechanism and mathematical model of EPLL framework,sparse coding model noise,image sparse representation,respectively,puts forward the EPLL image denoising algorithm based on sparse prior,sparse coding noise model combined with the EPLL framework,sparse dictionary training,denoising image block recovery weighting algorithm,such as key link,as an important innovation of this article,method of study for this paper.Finally will be similar with the method based on sparse dictionary K-the image representation and SVD denoising method,based on the framework of EPLL and Gaussian mixture model of fuzzy recovery image denoising algorithm and three-dimensional block matching(BM3D)algorithm is compared,from the viewpoints of both qualitative and quantitative verification this article proposed based on the EPLL framework with sparse prior recover blurred image denoising algorithm is effective.The results show that the proposed EPLL image denoising algorithm based on sparse prior qualitative analysis level,whether in the visual observation and peak signal to noise ratio,quantitative indexes such as structural similarity evaluation is better than that other popular image denoising restoration algorithms proposed in the paper.