改进的小波阈值去噪算法研究

Research on the improved wavelet threshold denoising algorithm

传统阈值函数包括硬阈值函数和软阈值函数,硬阈值函数存在不连续的问题,致使重构信号出现震荡,软阈值函数存在固定偏差,对重构信号影响较大。为了对传统阈值函数的不足之处进行改进,提出了一种新的阈值函数和阈值选取规则。通过选取最优小波基和分解层系数,利用MATLAB软件进行仿真实验,对比实验结果后发现,无论在信噪比（SNR）提升上还是均方根误差（RMES）上,新的阈值函数处理后的结果都优于传统方法,说明改进后的函数的处理效果得到进一步提升。

The traditional threshold function includes the hard threshold function and the soft threshold function,the problem of hard threshold function is discontinuous,it can cause a shock of the reconstructed signal,the soft threshold function has fixed deviation,it can have a big impact on the reconstructed signal.A new threshold function and threshold selection rule are proposed because of the Inadequacy of traditional threshold function.MATLAB simulation experiment was carried out by selecting the optimal wavelet base and decomposition layer coefficients,compare the data of the experiment,the new threshold function of the SNR and RMES are improved,the explanation can overcome the deficiency of traditional function to some extent.