基于Canny算子的小波域维纳滤波去噪方法

Denoising Method via Local Wiener Filtering in Wavelet Domain Based on Canny Operator

本文给出了一种利用Canny算子的小波域局部维纳滤波的图像去噪方法。先用二维小波对图像进行分解,在第一重维纳滤波时使用椭圆形方向窗估计信号的方差。在第二重维纳滤波时,使用Canny算子把图像分成边缘区域和非边缘区域,然后结合圆形窗估计信号的方差,再利用局部维纳滤波进行去噪。实验结果表明了本文方法的有效性。

This paper proposes a local wiener filtering image denoising method using Canny operator in wavelet domain. First, the noisy image is decomposed by a two - dimensional wavelet, in the first stage of wiener filtering, the signal variances of wavelet coefficients are estimated in the elliptic windows. In the second stage of wiener filtering, the canny op- erator is used to divide the image into edge and non - edge regions, and then combine the circle window to estimate the signal variance, finally the local wiener filtering is used to denoise the image. Experiment results show that the proposed algorithm is very efficient.