基于轮廓变换的自适应阈值图像去噪算法  

Adaptive threshold algorithm for image denoising based on Contourlet transform

作　　者：邓超[1] 刘泽乾[1] 侯满义[1] 杜跟随[2] 

机构地区：[1]空军航空大学作战指挥系,吉林长春130022 [2]空军航空大学训练部,吉林长春130022

出　　处：《电子设计工程》2013年第12期121-123,共3页Electronic Design Engineering

摘　　要：针对Contourlet变换子带系数相关的特性,对图像去噪的阈值进行了研究,提出了一种新的自适应阈值图像去噪算法。该方法首先对噪声图像进行Contourlet变换,得到不同尺度、不同方向上的子带系数,再根据子带系数的能量关系采用改进的Bayes Shrink自适应阈值对图像进行去噪处理。实验结果表明:这种方法提高了图像的PSNR值,增强了去噪后图像的视觉效果,而且更好地保留了图像的边缘信息。Concerning the character of the correlation coefficient of sub-bands in Contourlet transform, the threshold for image denoising was researched, and a new adaptive threshold algorithm was proposed which was used for image denoising. Firstly, the noise image was processed by Contourlet transform to get the coefficient of sub-bands in different scales and orientations, then according to the energy relation of the coefficient of sub-bands, the image was denoised base on the improved Bayes Shrink adaptive threshold. The experimental results demonstrate that： this method heightens PSNR value and enhances the visual effect of the denoised image, which can maintain the edges of the image better.

关 键 词：图像去噪 CONTOURLET变换 自适应阈值 BAYES Shrink方法 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] TN911.7[自动化与计算机技术—计算机科学与技术]