基于MMTD的改进的Bayes软阈值算法  

An Improved Bayes Soft Threshold Algorithm Based on MMTD

作　　者：李红丽 马耀锋 

机构地区：[1]郑州工程技术学院机电与车辆工程学院,郑州450044

出　　处：《中州大学学报》2017年第6期121-124,128,共5页Journal of Zhongzhou University

基　　金：河南省科技厅项目(172102210604);河南省高等学校重点科研项目(16B510008)

摘　　要：用贝叶斯阈值法来确定阈值,提出了基于MMTD的小波阈值图像去噪算法。采取阈值和小波系数之间的比值对系数较大的区域进行软阈值函数的萎缩操作;利用阈值周围的小波系数是否含有噪声的模糊性,对该模糊区域通过MMTD修改小波系数的保留程度。改进算法的仿真结果表明,基于MMTD的小波阈值图像去噪算法较典型算法相比,有更好的去除噪声效果,并较好地保持图像边缘细节。A new wavelet threshold image filtering algorithm combined MMTD is proposed based on Bayesian framework.The part where the value of wavelet coefficient is large,and processes through atrophying the value of wavelet coefficient,using the ratio of threshold to wavelet coefficient.The part,where the value of wavelet coefficient is fuzzy,is processed to determine the value of wavelet coefficient by using MMTD based on the fuzziness of wavelet coefficient around the threshold.Simulation results shows that the new algorithm can successfully remove noise and preserve the image edge details,compared with the classic algorithm.

关 键 词：图像去噪 中介真值程度的度量 贝叶斯阈值 硬阈值函数 软阈值函数 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]