一种基于相关方法的改进小波图像去噪算法

传统小波阈值法去噪方法在阈值选取上存在不确定性,为解决这一问题,提出采用小波相关方法,对图像噪声进行处理。在进行小波变换时,首先基于小波多方向框架理论进行图像分割,通过对多方向组合进行小波变换,最后对小波重构结果取中值得到去噪图像。该方法在处理具有方向特征的图像时,可以很好地消除Gibbs效应,与传统方法相比去噪结果具有明显改善。

An Improved Wavelet Correlation Algorithm for the Image De-noising

Traditional digital image de-noising method based on the wavelet thresholding method has been widely used,but uncertainty arises in defining the threshold. To solve the problem,the wavelet correlation method is applied in this paper. The image is divided by using the multi-directional frame theory,and the wavelet transformation is performed in the separated direction. The middle value is extracted to reconstruct the image form the inverse wavelet transformation in the multi-directions. Compared with traditional de-noising methods,the noise of images possessing directional character is better removed and the Gibbs effect is eliminated.