小波域上的图像降噪Wiener滤波器设计

摘要

小波阈值降噪技术利用小波变换表示信号的稀疏性质，使用对角形式的阈值滤波器达到信号降噪的目的，这个方法在很多信号空间上是近似最优的。但是对于任意信号，在MSE意义上最优的信号估计是Wiener滤波器。提出了一种小波域上的经验Wiener滤波器的设计方法，并应用于图像降噪。实验结果表明，此方法比一般小波阈值降噪法取得的效果要好得多。%Wavelet thresholding technology is using the sparse property of wavelet representation and diagonal filter for signal denoising. This method is nearly optimal in many signal spaces. But the Wiener filter is optimal for any space in the sense of MSE. A wavelet domain Wiener filter is developed in this paper. The design and the application in image denoising of this filter are proposed. The experiment result shows the Wiener filter is better than the standard wavelet thresholding method.