基于EMD-ICA的激电数据降噪处理方法

De-noising of IP data based on EMD-ICA

针对传统降噪方法处理激发极化法(激电法)测量数据的效果不理想的问题,对经验模态分解方法和独立分量分析技术进行了研究,提出了一种新的激电数据降噪方法。首先,采用经验模态分解方法将原始测量数据自适应分解为有限个固有模态函数,再根据其与激发信号的相关性选择固有模态函数构造虚拟噪声通道,最后利用独立分量分析技术提取多维混合数据中的激电信号。利用仿真信号和实际数据对该方法进行实验,对比普通滤波方法和小波阈值算法,结果表明该方法能有效提高激电数据的信噪比。

Aiming at traditional de-noising method may fail to remove noise from induced polarization （IP） data. This paper proposed a new de-noising method based on empirical mode decomposition （EMD） and independent component analysis （ICA）. Firstly, EMD can self-adaptively decompose data into finite intrinsic mode functions （IMFs）. Secondly, this paper constructed virtual noise channel according to the correlation between IMF and excitation signal. Finally, it used ICA to extract the effective IP signal. Eventually, this paper applied its method to de-noising in simulation and actual IP data. And by comparing the proposed method with those of traditional filter algorithm and wavelet threshold algorithm, the results verify the ef- fectiveness of the proposed method in de-noising of IP data.