赵继印,李先涛,赵静荣,郝志成,康朝海

• 1:吉林大学通信工程学院,东风设计研究院电气室,吉林大学通信工程学院,吉林大学通信工程学院,大庆石油学院电气信息工程学院 吉林长春　130025 ,湖北武汉　430056 ,吉林长春　130025 ,吉林长春　130025 ,黑龙江大庆　163318

摘要：

研究了噪声能量的估计方法:MAD法及其改进方法,空域相关估计法.在传统阈值法的基础上,提出半软阈值法.仿真结果表明:半软阈值去噪方法有效地抑制了噪声,提高了信噪比,同时保留了图像细节,获得了最小的均方差,具有一定的应用价值.

关键词： 图像信号;半软阈值法;噪声能量估计;去噪方法

Abstract: The estimation methods of noise energy are studied in this paper, which include MAD method and its improved null field correlation estimation method. Semi-soft threshold method is put forward based on classical threshold methods. The simulation results show that the semi-soft threshold denoising method reduces the noise effectively, improves the ratio of signal and noise, retains the detail of image and obtains the minimum mean square deviation. It finds some application in engineering.

Keywords: image signal;semi-soft threshold model;estimation of noise energy;denoising method

基金项目:

通讯作者:

Email:

参考文献：

• [1]　杨福生.小波变换的工程分析与应用[M].北京:科学出版社,1999.
• [2]　SandeepPG,AkbarMS,RichardGB.ImprovedWaveletDenoisingviaEmpiricalWienerFiltering[J].SPIE,1997,3169:389-399.
• [3]　DonohoDL.Denoisingbysoft thresholding[J].IEEETransonIT,1995,41(3):613-627.
• [4]　PanQuan,ZhangLei,DaiGuanzhong,etal.TwoDenoisingMethodsbyWaveletTransform[J].IEEETranSignalProcessing,1999,47(12):3401-3406.
• [5]　AndrewB,GaoHongye.WaveShrinkageFunctionsandThresholds[J].SPIE,1995,2569:270-281.