Abstract: Magnetotelluric signal has the characteristics of weak signal and wide frequency band. The quality of electromagnetic data obtained in the field is extremely reduced by strong human interference. In order to improve the signal-to-noise ratio of the collected data, it is necessary to adopt a reasonable denoising method to reduce the noise of the data. Wavelet transformation has the function of joint analysis of “ time-frequency", which can conduct multi-scale detailed analysis of signal. So it is very suitable for processing magnetotelluric data with complex frequency components. The wavelet transform principle combines the signal and noise characteristics of the field electromagnetic data, reasonable wavelet transformation parameters of parent wavelet, decomposition layer number and threshold are determined. Different combinations of parameters are used to handle different human noise interference. The signal-to-noise ratio of field electromagnetic data is effectively improved after wavelet transformation, which provides a high-quality time series for subsequent visual resistivity calculations.

Key words: magnetotelluricsignal (MT) , strong human interference , wavelet transform , signal-to-noise ratio (SNR)