基于多小波变换的图像去噪  

IMAGE DE-NOSING METHOD BASED ON MULTI-WAVELET TRANSFORM

作　　者：董卫军[1] 樊养余[1] 周明全[2] 

机构地区：[1]西北工业大学电子信息学院,陕西西安710072 [2]西北大学信息科学与技术学院,陕西西安710069

出　　处：《计算机应用与软件》2007年第12期32-33,100,共3页Computer Applications and Software

基　　金：国家自然科学基金(60573179);陕西省科技计划工业科技攻关项目(2004k05-G24)的资助

摘　　要：小波分析是目前国际上最新的时间频率分析工具,是信号去噪的强有力处理工具。小波变换可以将交织在一起的混合信号分解成不同频率的块信号。多小波所拥有的对称性、正交性、有限支撑等重要特性弥补了单小波的不足。提出了一种新的图像去噪方法,该方法以多小波变换为基础。实验证明,去噪效果良好。Wavelet analysis ,the latest time-frequency analysis tool internationally,is a powerful tool for image de-nosing. Through wavelet transform, signals are separated into block signal with different frequency. Multi-wavelet has some very important properties such as orthogonality,symmetry,and finite support,which make up for the shortcoming of scalar wavelet. A new image de-nosing method based on multi-wavelet transform is proposed. The experimental results show that the new method can achieve good de-nosing effect.

关 键 词：多小波 图像去噪 小波变换 

分 类 号：TN919.81[电子电信—通信与信息系统]