基于多图像的自适应小波去噪

Wavelet Adaptive Denoising Based on Multiple Images

为有效地去除图像噪声,提出了采用2幅或者多幅输入图像的去噪算法.该算法通过对2幅或者多幅被不同等级的噪声所污染的图像进行正交小波变换,对变换后的系数进行加权运算,然后采用自适应于尺度和小波子带大小的自适应阈值方法进行去噪,以突出图像的特征,并减少噪声的影响.试验结果表明,与其他几种去噪方法相比,本算法具有良好的视觉效果,并且峰值信噪比也有较大幅度的提高.

A denoising method based on two or multiple input images is proposed in the paper in order to remove noise effectively.Two or multiple images contaminated by different levels of noise are decomposed by a 2D discrete wavelet transformation,then the weighed computation is applied after the transformation,and finally the threshold which is adaptive to the scale and the size of the sub-band is adopted in the denoising.Experimental results show that compared with the other algorithms,the algorithm obtains better visual effect and the PSNR is also greatly enhanced.