一种基于新型小波包阈值的图像去噪方法  

A Method for Image Denoising Based on New Wavelet Packet Threshold

作　　者：胡波[1] 陈恳[1] 徐建瑜[1] 

机构地区：[1]宁波大学信息科学与工程学院,浙江宁波315211

出　　处：《宁波大学学报（理工版）》2009年第4期454-458,共5页Journal of Ningbo University：Natural Science and Engineering Edition

基　　金：浙江省教育厅科研计划项目(20070869)

摘　　要：提出了一种基于小波包理论去除图像噪声的方法,用小波包把图像分解为高频分量和低频分量,根据高频分量估计噪声的标准差,并利用该标准差以及Birge-Massart惩罚函数计算阈值.鉴于传统软硬阈值的缺陷,采用一种新型阈值量化方法,用三次多项式在硬阈值的基础上插值,使新的阈值函数保持了连续性和可导性.通过这种方法既消除了图像的振铃现象,又保留了细节成分.实验表明:与传统方法相比,新方法使图像视觉效果和峰值信噪比均获得提高.An image denosing method based on wavelet packet theory is proposed in this paper.Wavelet packet is first used to decompose the image into the low and high frequency components.The standard deviation of the noise is estimated through high frequency components.Then,the estimated standard deviation is applied together with Birge-Massart penalized function to calculate the threshold.To remedy drawbacks identified in the traditional soft and hard threshold for image denoising,this paper presents a new threshold quantization method,which applies cubic polynomial interpolation to a hard threshold to achieve the continuity and differentiability for the new threshold function.The application of this method not only removes the ring effect from the image,but also keeps the image details.The experiment results show that,in comparison with the conventional denoising approaches,both the visual effect and the PSNR have been further improved using the proposed method.

关 键 词：图像去噪 小波包 惩罚函数 新型阈值 三次多项式 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]