Abstract: Aiming at the issue of threshold function setting by using wavelet transform signal denoising, an improved method of threshold function based on the classical method of soft and hard threshold function denoising was presented. After this improved threshold function method and the maximal overlap discrete wavelet packet transform were combined, a wavelet denoising method of improved threshold function was obtained. Simulation results obtained from Matlab software indicated that this method can improve signal-to-noise ratio(SNR) of the reconstructed signal, remove the noise effectively, and reserve the main characteristics of the original signals. So, it is better than the traditional methods. This method has a practical significance in application of stock denoising.

Key words: acoustics , improved threshold function, maximal overlap discrete wavelet packet transform , wavelet denoising , noise filtering

摘要： 针对用小波变换进行信号去噪的阈值函数设定问题，在传统软、硬阈值函数去噪的基础上，提出一种改进的阈值函数方法，并与极大重叠离散小波包变换相结合，从而得到一种改进阈值函数的小波去噪方法。[Matlab]仿真结果表明：去噪方法提高了重构信号的信噪比，有效除去噪声，且保留原始信号的细节特征，是一种较好的信号消噪方法，在股票去噪中具有较高的实用价值。

关键词: 声学, 改进阈值函数, 极大重叠离散小波包变换, 小波去噪, 噪声滤波