均匀设计在小波图像去噪阈值选取中的应用  

Uniform Design of Threshold Parameter in Threshold-basedImage De-noising by Wavelet Transform

作　　者：黄明辉[1] 朱维彰[1] 

机构地区：[1]杭州电子工业学院计算机分院,浙江杭州310018

出　　处：《杭州电子工业学院学报》2004年第1期79-82,共4页Journal of Hangzhou Institute of Electronic Engineering

摘　　要：基于阈值的小波图像去噪的核心环节是阈值的选取。阈值δ=cσ被证明是很有效的方法。在硬阈值δ=cσ的图像去噪方法中,其去噪效果的关键为适当确定阈值大小,即参数c值的选取。但一般来说最佳的参数c会随图像的不同和迭加噪声的强度而变化。通过均匀设计可以估计出c在不同强度的噪声污染情况下的最佳取值区间,从而达到更有效的图像去噪目地。故利用均匀设计方法进行阈值参数c的选取,可以取得了很好的效果。The choosing of threshold is a key in threshold-based image de-noising by wavelet transform.The threshold δ=cσ is proved to be efficient in threshold-based image de-noising by wavelet transform.Generally speaking,the optimal c varies with the images and the intension of the added noise. The best fetching value block of the constant c can be estimated appearing in the images with the intension of the added noise by uniform design method. Thus it can reduce noise more effectivelly. This paper intruduces uniform design method used in the constant c selection .It has been shown is an effective method.

关 键 词：均匀设计 小波变换 图像去噪 阈值选取 阈值萎缩方法 

分 类 号：TN911.73[电子电信—通信与信息系统]