摘要： 提出利用小波模极大值进行图像消噪方法，对含噪声图像进行离散平稳小波变换和噪声标准差的估计，在Bayes-shrink阈值计算的基础上，得到消噪的阈值计算公式。对各尺度各子带的小波系数模极大值进行判断，获得由图像边缘产生的小波系数，使用自适应多阈值的方法在小波各尺度、各子带萎缩非图像边缘产生的小波系数。经平稳小波逆变换得到消噪后的图像。实验结果表明，与以前消噪方法相比，该方法具有更好的效果。

关键词: 平稳小波变换, 模极大值, 萎缩, 消噪, 清晰度

Abstract: This paper proposes using wavelet modulus maxima de-noising method. Image containing noise is took discrete stationary wavelet transform, the noise standard deviation is estimated, changes in the form of Bayes-shrink threshold, and gets de-noising threshold formula. The paper judges the wavelet coefficients modulus maxima generated by the edge image in each scale wavelet coefficients of each sub-band, uses the adaptive multi-threshold to atrophy wavelet coefficients for non-image edge in each wavelet sub-band of all different scales. The stationary wavelet inverse transform of the image is performed after de-noising. Experimental results show that the proposed de-noising method has better effect.

Key words: stationary wavelet transform, modulus maxima, shrink, de-noising, sharpness

中图分类号: 

• TP393.41