Suppression method of MT interference noise based on marginal spectrum and blind source separation

As magnetotelluric exploration has two inherent disadvantages of "weak surface observation signal" and "random excitation of field source", it is extremely susceptible to various kinds of interference noises which have serious effects on data processing and inversion interpretation in the later stage. Therefore, it is urgent to suppress the interference noise in MT signal. In this paper, blind source separation is performed for suppression of magnetotelluric interference noise. Firstly, the magnetotelluric data with interference noise is transformed by HHT, and then the marginal spectrum is used to determine the number of source signals in order to do blind source separation. Then the IMF components by EMD processing and the original signals are combined into a signal group to conduct independent component analysis. At last, a dynamic factor processing is used in order to avoid the amplitude displacement in the extracted signal. After the processing of the simulation signal and actual magnetotelluric signal, compared with wavelet de-noising method, it is found that the method proposed by this paper can suppress the magnetotelluric interference noise more effectively.