Second Order Total Generalized Variation Image Denoising A Algorithm With Alternating Direction Method Of Multipliers

Image is the main way to obtain information for human being.It is easy to be interfered and produced noise during the process of generation,transmission and storage.The existence of noise reduces image quality and affects people's understanding and processing of image information.The purposes of image denoising are to reduce noise in image and provide convenience for the subsequent processing of the image.Image denoising integrates the relevant knowledge of computer science and technology,image processing and mathematical analysis methods,and has been widely applied in many fields such as medical imaging,military science,satellite remote sensing,face recognition and artificial intelligence.This paper proposes an improved second-order total generalized variation denoising algorithm for the "staircase effect" problem of total variation image denoising algorithm.This method is proposed on the basis of the total generalized variational gorithm.By introducing three auxiliary variables x,y and z,the original unconstrained problem is rewritten into a constrained problem model.The model is solved by the alternating direction multiplier method(referred to ADMM),In this article,the ADMM algorithm is slightly improved to speed up the experiment.The experimental results show that compared with the classical total variation denoising model,The new model is more excellent in denoising effect,reduce the “staircase effect” to a certain extent,and obtain better PSNR in experimental data.