一种小波自适应比例萎缩去噪改进算法

摘要

作为小波应用于图像去噪的一种重要的方法,小波比例萎缩法具有很强的局部自适应能力,它的去噪效果与Donoho 的收缩法相比有了很大的提高.但是比例萎缩法有其自身缺陷,重建后的图像容易出现“波纹”状噪声.从产生这种噪声的原理入手,提出了一种对方差估计进行改进的算法,很好地克服了比例萎缩法中出现的这种噪声,提高了图像的去噪质量.%As an important method applied to image de-noising in wavelet field, adaptive proportional-shrinking algorithm has a strong local adaptive capacity in wavelet domain. Compared with Donoho's shrinking method, its effect has been greatly improved. But the proportional-shrinking algorithm has defects and the image is prone to the ripple-noise after reconstruction. Starting from the principle to produce this noise, an algorithm which improves the variance estimation is proposed in this paper. The ripple-noise which appeared in proportional-shrinking algorithm is well overcome and the quality of image is greatly improved.