基于无下采样的正交小波变换的阈值去噪方法  

Threhold De-noising based on Orthogonal Wavelet Transform without Down-sampling

作　　者：董勇[1,2] 李梦霞[1] 陈忠[1] 

机构地区：[1]长江大学信息与数学学院,湖北荆州434023 [2]武汉理工大学理学院,湖北武汉430070

出　　处：《长江大学学报（自科版）（上旬）》2009年第1期10-12,共3页JOURNAL OF YANGTZE UNIVERSITY (NATURAL SCIENCE EDITION) SCI ＆ ENG

基　　金：国家自然科学基金项目(40572078/D0206);教育部重点实验室开放基金项目(K200609);石油科技中青年创新基金项目(2002f70104);湖北省教育厅(A类)重点项目(D200512001)

摘　　要：现行的小波阈值去噪算法中,都包含采样的操作,但下采样导致信号所含信息出现丢失,上采样则引入新的畸变。为克服这些不足之处,提出了基于无下采样的正交小波变换的阈值去噪方法,并在Matlab6.5环境下给出了具体的算法。并进行了数值试验。试验结果表明,取消下采样的操作,可以提高信噪比,减小最小均方误差,使去噪效果得到了明显改善。In the current popular methods,based on wavelet threshold de-noising,there are sampling operations.But down-sampling operatin produced the loss of the information contained in the signal,and up-sampling produced extra aberration of the signal.To conquer the shortcoming mentioned above,the paper gives a method based on orthogonal wavelet transform without down-sampling,and discribes material algorithm under the condition of Matlab 6.5.Numerical simulation is also carried out,the results indicate that the de-noising effects are improved.

关 键 词：正交小波变换 下采样 阈值 去噪 信噪比 最小均方误差 

分 类 号：TN911.72[电子电信—通信与信息系统]