基于自由分布式FDR假设检验阈值算法的信号去噪

Signal denoising based on threshold algorithm of free distributed FDR hypotheses testing

在小波阈值去噪算法中存在的主要问题是阈值的设定和阈值函数的选取。Donoho提出的通用阈值在实际的应用中效果并不十分理想。利用统计学中的自由分布式错误发现率假设检验方法来设定阈值,使得该阈值不依赖于信号的长度,并在软阈值函数和硬阈值函数的基础上提出一种改进型阈值函数。仿真结果表明,与Donoho提出的小波阈值去噪算法相比,该算法具有更好的去噪性能。

The main problem of wavelet threshold denoising algorithm is threshold setting and threshold function selection. The universal threshold proposed by Donoho is not very ideal in practice. According to the method of free distributed False Discovery Rate hypotheses testing in statistics, the threshold is set and it does not depend on the length of signal. Moreover, a modified threshold function is proposed on the basis of soft threshold function and hard threshold function. The simulation results show that the proposed algorithm is more effective than that of wavelet threshold denoising algorithm proposed by Donoho.