基于小波变换的边缘保留图像去噪改进算法  

Improved Edge-preserving Image Denoising Algorithm Based on Wavelet Transforms

作　　者：刘平[1] 王磊[1] 郭中华[1] 李树庆[1] 

机构地区：[1]宁夏大学物理电气信息学院,宁夏银川750001

出　　处：《电视技术》2014年第5期13-15,30,共4页Video Engineering

基　　金：宁夏回族自治区自然科学基金项目(NZ1103)

摘　　要：针对在图像去噪过程中,如何有效地保留图像边缘等重要特征信息的问题,提出了一种基于小波变换的图像去噪改进算法。对图像进行多尺度小波分解,将各子带小波系数进行自适应阈值化处理,边缘成分的阈值由子带阈值和给定的相关权重计算得到,从而有效保留图像边缘信息。分别对Tracy和Building图像进行处理,实验结果表明,与BayesShrink等4种传统方法相比较,改进算法不仅可以有效去除不同程度的加性高斯白噪声,很好地保留图像边缘等重要特征信息,而且具有较高的峰值信噪比。An improved image denoising method based on wavelet transforms is proposed to resolve the problems of how to preserve important features effectively such as edges of the image during the image denoising process. The image is decomposed by Multi-scale wavelet, and then wavelet coefficients of each sub-band are processed by the adaptive threshold. The thresholds of the edges are calculated by the sub-band thresholds and the given weights, which can effectively preserve the edge information. The images Traey and Building are processed respectively and the experimental results are compared with other four traditional approaches such as BayesShrink, demonstrating that the proposed method can not only remove different levels of Additive-White

关 键 词：图像去噪 边缘 小波变换 自适应阈值 峰值信噪比 

分 类 号：TN911.73[电子电信—通信与信息系统] TP315[电子电信—信息与通信工程]