A Partial Differential ANRDPM Image Denoising Model Based on A New Anti-Noise Coefficient and Reverse Diffusion Idea  

作　　者：Yuze Chen Xianchun Zhou Siqi Lu Binxin Tang Mengnan Lv Zhiting Du 

机构地区：[1]School of Electronic and Information Engineering,Nanjing University of Information Science and Technology,Nanjing 210044,China [2]School of Artificial Intelligence(School of Future Technology),Nanjing University of Information Science and Technology,Nanjing 210044,China

出　　处：《Instrumentation》2024年第4期21-34,共14页仪器仪表学报(英文版)

基　　金：funded by National Nature Science Foundation of China,grant number 61302188.

摘　　要：To overcome the problem of insufficient expression of fine texture when gradient mode is used as an image feature extraction operator in traditional PM model,which leads to excessive diffusion in these fine texture regions and texture ambiguity,this paper proposes ANRDPM(Anti-noise and Reverse Diffusion PM model)noise reduction model based on the new anti-noise coefficient and reverse diffusion concept.In this model,the meter gradient operator is used as the image feature extractor to solve the shortage of the traditional gradient operator in the ability to express details.Secondly,a new anti-noise coefficient based on Gaussian curvature and noise intensity is proposed to solve the problem that the meter gradient operator is allergic to large noise points.In addition,a reverse diffusion filter based on a local variance of residuals is introduced to enhance the smoothed texture information in the image.Finally,the new model is discretized by a finite difference algorithm,and simulation results show that the proposed ANRDPM model not only performs well in smoothing image noise,but also effectively protects image texture information and structural integrity.

关 键 词：image noise reduction partial differential equation new anti-noise coefficient anisotropy total variation 

分 类 号：O17[理学—数学]