基于多级中值滤波的小波去噪方法  

Wavelet denoising based on multistage median filtering

作　　者：贺长伟[1] 刘英霞[1] 任文杰[1] 王欣[1] 

机构地区：[1]山东大学信息科学与工程学院

出　　处：《计算机应用》2007年第9期2117-2119,2125,共4页journal of Computer Applications

基　　金：山东省自然科学基金资助项目(Z2005G02)

摘　　要：经典图像去噪方法中,多级中值滤波有良好的保护细节的特性,近年来在小波域中对信号的处理,能使噪声抑制更加有效。结合两者的特点,提出了基于多级中值滤波的改进算法,利用最大最小中值之差判断平坦及边缘区域,实现了在小波域中平滑噪声的同时还可以保护图像细节不受损失。实验表明该方法与Donoho的软门限方法相比较,可以得到更好的去噪效果。In traditional image denoising methods, multistage median filter can preserve the details effectively. Recently, the wavelet transform has been used for image denoising. Then to combine the good qualities of them, an improved way based on the multistage median filter was proposed and it can differentiate the flat and edge regions to protect the details and smooth the noise. Experimental results show that the proposed algorithm can get better denoising result than the soft thresholding proposed by Donoho.

关 键 词：小波变换 图像去噪 多级中值滤波 最佳估计 

分 类 号：TN911.73[电子电信—通信与信息系统]