结合Shearlet与Bayesian MAP估计的图像去噪  

Image Denoising Through Combination of Shearlet Transform and Bayesian MAP Estimation

作　　者：王飞[1] 史彩成[2] 朱士祥[1] 

机构地区：[1]中国人民解放军93575部队 [2]北京理工大学生命学院

出　　处：《电光与控制》2015年第4期50-53,共4页Electronics Optics & Control

摘　　要：针对小波变换及Contourlet变换在图像去噪应用中的不足,提出一种结合Shearlet与Bayesian MAP估计的图像去噪算法。首先对含噪图像进行Shearlet分解,之后根据Bayesian MAP准则对分解后的各子带信号进行估计,最后通过重构得到去噪图像。实验表明,相对于其他去噪算法,该方法很好地保留了图像的细节,取得了更好的视觉效果。Considering the shortcomings of Wavelet and Contourlet when applied to image denoising,we proposed a new method based on Shearlet transform and Bayesian MAP estimation. Firstly,the source images were decomposed into several subbands using Shearlet. Then,Bayesian MAP estimation was adopted to estimate these multi-direction subbands. Finally,the denoised image was obtained by performing the inverse Shearlet transform on the coefficients. The experimental results show that： Compared with other denoising method,this method can contains more details,and get better visual effects.

关 键 词：图像去噪 图像重构 SHEARLET变换 BAYESIAN MAP 

分 类 号：V271.4[航空宇航科学与技术—飞行器设计] TN911[电子电信—通信与信息系统]