摘要
根据小波变换和噪声信号的能量分布特性,提出了一种先用小波变换对含噪图像进行多尺度分解,求出各尺度小波变换高频系数的噪声方差和阈值,利用各尺度的阈值对高频系数进行处理,然后利用小波变换系数重构图像,实现图像降噪的方法;实验结果说明该方法既可以有效地降低噪声,又可以较好地保持图像细节。
According to the characteristic of energy distribution of wavelet transform and noise,a method of image noise reduction is proposed. Firstly, noised image is decomposed by wavelet transform with multi-scale. Secondly, the variance of noise and threshold in high frequency coefficients of wavelet transform with different scale are figured out. And the coefficients are dealt with by using different threshold. Finally, reconstructed image can be obtained by using inverse wavelet transform for all coefficients. Experimental results prove that by using this method,image noise can be reduced effectively and image details can be preserved a lot.
出处
《机械工程与自动化》
2008年第5期32-34,共3页
Mechanical Engineering & Automation
关键词
小波变换
阈值
图像降噪
wavelet transform
threshold
image noise reduction