基于小波和Contourlet的改进的图像复原算法

不同的数学变换工具能够有效表示图像的不同细节结构,小波变换能够有效表示图像中的奇异点,而con-tourlet变换能够有效表示奇异线,为了更好地利用不同变换工具的优势,文中提出一种基于小波和Contourlet的改进的图像复原算法。算法首先分别应用不同的小波基和不同的Contourlet基,基于正则化方法求解出复原图像;然后,将经过不同的滤波器组得到的复原图像通过加权平均的方式融合,得到一幅效果较好的恢复图像。实验结果表明,加权平均之后的图像相比使用单一滤波器的复原图像,其改善的信噪比提高0.1～0.5dB。

Image restoration algorithm based on wavelet and contourlet

Because different mathematical transformation tool can effectively show different details structure of image,an algorithm combining wavelet and contourlet is proposed in order to utilize the various advantages of the transform.Firstly,the algorithm uses various wavelet basis and various contourlet basis to restore the images based on regularities algorithm.Second,the restored images based on contourlet are average weighted,and then find the edge and texture using gradient operator.The restored images based on wavelet are averaged too,minus the edge and texture and so obtain the smooth region.At last,it fuses the smooth,the edge,and the texture to one good quality image.Experiments shows that the improved-signal-noise-ratio of restored image by fused image is 0.1～0.5 dB better than that by any single transform.