基于BEMD与自适应维纳滤波的图像降噪

Image denoising method based on BEMD and adaptive Wiener filter

自适应维纳滤波器是一种经典的线性降噪滤波器,较其他线性滤波器能够更好地解决边界模糊的问题。然而由于含噪图像的噪声主要集中于它的高频部分,而图像的低频部分所含有的噪声较高频部分则小很多。自适应维纳滤波算法对图像中所有频率成份都不加区分地进行滤波降噪处理,因而它不能得到更为令人满意的结果。提出了一种将二维经验模态分解和自适应维纳滤波相结合的图像去噪方法,通过将图像分解为不同频率成份的子图像并对各子图像采用不同的降噪处理,从而更好地对含噪图像进行降噪。实验结果表明,算法相对于自适应维纳滤波算法降噪效果更好。

Adaptive wiener filter, a classical linear denoising filter, can solve the problem of fuzzy boundary more effectively than other linear filter. However, the noise mainly exists in high frequency part of a picture, and much less in the part of low frequency, while adaptive wiener filter tends to deal with the noise without considering the different frequency componets of the picture, therefore, it can not get a satisfactory result. A new method combining BEMD method and adaptive wiener filter is proposed in this paper, which can denoise better through decomposing a picture into different part accoding to the different frequency and then denoising seperately. The result of the experiment in this paper shows that the proposed method performs much better than adaptive wiener filter.