Abstract: In order to enhance effect of adaptive wavelet threshold de-noising method, a new wavelet threshold de-noising method combining wavelet packet decomposing based on maximum energy matching and fast sample entropy is presented. According to feature of each layer wavelet coefficient, wavelet packet base is selected adaptively based on maximum energy matching criterion to decompose signal with noises and then the wavelet coefficient under maximum scale is processed with threshold and is restructured to get noise signal. Fast sample entropy algorithm is used to compute sample entropy of noise and the threshold is adjusted dynamically to make sample entropy of noise maximum and obtain optimal de-noising effect. The results by example show that the method presented in this paper has better de-noising effect compared with traditional wavelet threshold de-noising.

Key words: maximum energy matching criterion, fast sample entropy, wavelet threshold de-noising

摘要： 为提高自适应小波阈值降噪方法的效果，提出一种结合最大能量匹配的小波包分解和快速样本熵的小波阈值降噪方法。根据各层小波系数特点并以最大能量匹配准则自适应选择小波包基对含噪信号进行分解，对最大尺度下的小波系数阈值化后重构得到噪声信号，采用快速样本熵算法计算噪声信号样本熵，动态调整阈值使噪声信号样本熵最大而获得最佳的降噪效果。应用实例表明：该方法相比传统的小波阈值降噪方法具有更好的降噪效果。

关键词: 最大能量匹配准则, 快速样本熵, 小波阈值降噪