小波系数局部特征的自适应图像降噪算法  

Adaptive Algorithm for Image Denoising Based on Local Characteristic of Wavelet Coefficient

作　　者：吕俊白[1] 

机构地区：[1]华侨大学计算机科学与技术学院,福建泉州362021

出　　处：《华侨大学学报（自然科学版）》2010年第6期636-640,共5页Journal of Huaqiao University(Natural Science)

摘　　要：在Visu Shrink和基于Bayes准则的Bayes Shrink去噪方法的基础上,提出一种基于小波系数局部特征的自适应图像降噪算法.该算法从含噪图像的HH1子带估算噪声信号的标准差,并据此优化小波分解所需的级数;然后,根据小波系数的局部特征,自适应地选择不同子带不同方向上的最佳阈值,运用软阈值函数对图像进行降噪.与传统方法相比,该方法不仅提高图像的峰值信噪比,使图像更清晰,而且具有实现简单、运算速度快的特点.Based on the Visu Shrink and the Bayes Shrink derived in a Bayesian framework,a new adaptive algorithm for image denoising based on local characteristic of wavelet coefficient is proposed.First,the noise standard deviation is estimated from the subband HH1 to optimize the scale in the wavelet decomposition,then the optimal threshold for different subbands and orientations is determined according to the local characteristics.The image denoising is made by using soft-thresholding function.Comparing with traditional denoising algorithm,this algorithm can improve the peak signal to noise ratio（PSNR） more effectively and also makes denoised image more clearly,it can compute fast with a simple implementation.

关 键 词：图像降噪 整数提升 小波变换 分解级数 自适应阈值 峰值信噪比 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]