基于Donoho模型和高阶统计理论的小波消噪算法研究及其应用  

Study and application of wavelet denoising algorithm based on Donoho denoising module and HOS

作　　者：蒋云霄[1] 杨俊安[1] 钟子发[1] 沈辉[1] 

机构地区：[1]解放军电子工程学院装备技术研究中心,安徽合肥230037

出　　处：《电路与系统学报》2007年第1期11-14,共4页Journal of Circuits and Systems

基　　金：国家自然科学基金资助项目(60171029)

摘　　要：本文结合经典Donoho小波消噪模型,提出了基于高阶统计理论的小波消噪算法。算法在小波分解各尺度引入了高阶统计理论中的四阶累积量方法,从信号和噪声的统计特性出发进行信噪分离。仿真结果验证了本文算法的性能。算法能满足无线电监测系统中对信号进行后续处理的需求。Based on the Donoho denoising model, the paper presents a wavelet-based denoising algorithm with the fourth-order cumulant in Higher Order Statistics （HOS）. The proposed algorithm introduces the fourth-order analysis method in HOS across all wavelet scales, and detects the signal coefficients from the noise by their different statistics characters. Simulation results show that the proposed algorithm has a remarkable superior ability to denoise. The algorithm can satisfy the demand of signal processing in radio inspection system.

关 键 词：小波变换 四阶累计量 消噪 无线电监测 

分 类 号：TN958[电子电信—信号与信息处理]