基于自适应多分辨率奇异值分解的大地电磁数据处理

针对强电磁干扰极易掩盖微弱的大地电磁有用信号，本文结合奇异值分解在去噪方面的优越性，提出基于自适应多分辨率奇异值分解（Adaptive Multi-Resolution Singular Value Decomposition，AMRSVD）的大地电磁数据处理方法.首先对大地电磁数据构建Hankel矩阵，利用MRSVD得到不同分辨率的近似信号和细节信号；然后选用近似信号和细节信号的标准差差值，对大地电磁数据进行信噪辨识；接着结合MRSVD和相邻细节信号的标准差差值，提出先验信息未知情况下的AMRSVD法；最后对辨识出的强干扰运用AMRSVD去除噪声，重构有用信号.实验结果表明，该方法的处理效率高，能有效分离出相关性较强的噪声，时间序列和视电阻率-相位曲线均得到有效改善.

Magnetotelluric data processing based on adaptive multi-resolution singular value decomposition

In view of the strong electromagnetic interference can easily cover up the weak useful magnetotelluric signal, combined with the advantages of singular value decomposition in denoising, a new magnetotelluric data processing method based on adaptive multi-resolution singular value decomposition (AMRSVD) is proposed. First, the Hankel matrix is constructed for the magnetotelluric data, the approximate signal and detail signal with resolution can be obtained by MRSVD. Then, the signal-noise identification of magnetotelluric data is carried out by using the standard deviation difference between the approximate signal and the detail signal. Then, combined with MRSVD and adjacent standard deviation of detail signal, AMRSVD method is proposed when prior information is unknown. Finally, the AMRSVD method is used to remove the noise for the identified strong interference and reconstruct the useful signal. The experimental results show that this method has high data processing efficiency, and it can effectively remove the noise with strong correlation, the time series and apparent resistivity-phase curves are improved obviously.