基于一种渐进半软阈值函数的小波去噪  

Wavelet Denoising Based on an Asymptotic Semisoft Thresholding Function

作　　者：吴安全 沈长圣[1] 肖金标[1] 孙小菡[1] 

机构地区：[1]东南大学电子器件和系统可靠性研究中心,南京210096

出　　处：《电子器件》2017年第2期396-399,共4页Chinese Journal of Electron Devices

基　　金：江苏高校品牌专业建设工程项目

摘　　要：针对硬软阈值函数的不足,提出了一种渐进半软阈值函数,不仅弥补了硬阈值函数不连续性的缺陷,同时还克服软阈值函数重构信号后存在恒定偏差的缺陷,使得估计的小波系数逐渐接近于真实小波系数。通过对含噪Doppler信号的MATLAB仿真实验结果表明,与传统阈值函数和几种改进阈值函数相比,渐进半软阈值函数在去噪性能上能够获得更优的信噪比(SNR=21.443 1 dB)和均方根误差(RMSE=0.343 3),更适合于实际应用。For defect of the hard and soft threshold function,a semisoft threshold function is proposed.The function not only compensates for that the defect of the hard threshold function is not continuous,but also overcome the problem of that the soft threshold function exists constant bias in reconstructed signal.The semisoft threshold function makes the estimated wavelet coefficients of the wavelet coefficients gradually closer to the real.The results of MATLAB simulations show that Doppler affected by noise can get better SNR which is 21.443 1 dB and RMSE which is 0.343 3 on de-noising performance,compared with the traditional threshold functions and several improved threshold function.Therefore,it is more suitable for practical applications.

关 键 词：信号处理 小波变换 阈值去噪 信噪比 均方根误差 

分 类 号：TN911.4[电子电信—通信与信息系统]