基于小波变换的混合阈值去噪的方法研究  

Mixed threshold denoising based on wavelet transform

作　　者：邓禹[1] 熊静琪[1] 范显峰[1] 

机构地区：[1]电子科技大学机械电子工程学院,四川成都610054

出　　处：《光通信研究》2008年第3期67-69,共3页Study on Optical Communications

基　　金：国家自然科学基金资助项目(60672165)

摘　　要：基于小波变换单一阈值去噪方法由于其采用固定模型进行阈值估计,在某种情况下去噪效果较好,但是在特定工程中去噪效果常常不能令人满意。文章提出一种基于小波变换的混合阈值去噪方法,对不同阈值加入权重,通过目标函数优化权重,计算出较理想的阈值。仿真结果表明,与传统方法相比,该方法有较好的去噪效果并能够保留细节特征。Single threshold denoising based on wavelet transform is effective in certain cases as fixed models are used for threshold estimation, hut it is often unsatisfactory in some specific projects. This paper presents a wavelet transform-based mixed threshold denoising method, in which weight coefficients are added to different thresholds and fairly ideal threshold are obtained by optimizing the weight for the object function. The simulation results show that compared with the conventional methods, this method better in denoising effect and maintains the detailed features.

关 键 词：小波变换 混合阂值 去噪 

分 类 号：TN911.72[电子电信—通信与信息系统]