基于数学形态学与小波阈值组合滤波算法的大地电磁噪声压制

Noise Suppression of Magnetotclluric Sounding Data Based on Mathematical Morphology Combined with Wavelet Threshold

针对实测大地电磁信号资料中常出现的强噪声干扰问题,提出一种基于数学形态滤波和小波阈值滤波结合的大地电磁噪声压制方法。首先,利用自适应多尺度形态滤波算法对含噪信号进行初次滤波,以消除脉冲类干扰。然后,对已处理的信号进行小波阈值滤波提取大尺度强噪声轮廓。最后,剔除强噪声干扰重构大地电磁信号。通过仿真实验对比,在不同强度噪声干扰背景下,所提方法的性能优于普通小波阈值方法,能更多地保留大地电磁原始信号的细节特征。实测资料处理结果表明,该方法能有效地抑制大尺度强噪声干扰和基线偏移的影响,改善视电阻率曲线质量,具有较好的应用前景。

In order to suppress strong noise interferences which often exist in measured magnetotelluric(MT)data,the method of magnetotelluric noise suppression based on the combination of mathematical morphology filtering and wavelet threshold filtering was proposed.First,the adaptive multi-scale morphological filter was used to suppress impulsive type interferences.Then,the processed signals were filtered by wavelet threshold filter to get the contours of large scale noises.Finally,the strong noises were removed to reconstruct the magnetotelluric signals.Through the comparison of simulation experiments,it proved that the performance of the proposed method is better than that of the ordinary wavelet threshold method under the background of noise interferences of different intensities,and the more detailed features of the original magnetotelluric signals can be retained.The experimental results showed that this method can effectively suppress the influence of large-scale strong noises and baseline drift,and improve the quality of the apparent resistivity curves,thus the proposed method has a good prospect of application.