基于经验模态分解(EMD)的小波熵阈值地震信号去噪

Wavelet Entropy Threshold Seismic Signal Denoising Based on Empirical Mode Decomposition(EMD)

针对EMD阈值去噪算法中阈值由经验选取以及无法有效区分各固有模态函数上有用信息的不足,本文对各固有模态函数进行小波变换,对各层小波系数进行相关处理,以突出有效信息,抑制噪声;将细节系数的有效信号和突变点置零并等分为若干区间,选取小波熵最大子区间的高频小波系数平均值作为噪声方差计算得到阈值。该阈值选取方法依据小波熵的特点,自适应地根据对应尺度上信号自身的能量特征确定该尺度阈值。将该算法应用于仿真信号和实际地震信号去噪,结果表明该方法优于基于EMD的小波阈值去噪,在提高去噪效果的同时,也更好地保护有效信号。

In view of the threshold of empirical mode decomposition （EMD） threshold denoising algorithm selected by experience, and its inability to effectively distinguish useful information in the intrinsic mode function, the authors use the wavelet transform of the intrinsic mode function to process each layer of wavelet coefficients so as to highlight the effective information and suppress noise. After the effective signals and the mutation points of the detail coefficients are set to zero, the new detail coefficients are equally divided into several intervals. We select the wavelet entropy as the high frequency wavelet coefficients of the architectural interval average noise variance, and calculated the threshold value. The threshold selection method is based on the characteristics of the wavelet entropy, adapted to the energy characteristics of the corresponding scale signal itself in determination of the scale of the threshold. The algorithm is applied to the simulation signals and real seismic signal denoising. The results show that the method is better than that of the wavelet threshold denoising based on EMD; at the same time it can better protect the effective signals.