CNN-LSTM-based magnetotelluric signal noise suppression method and system

The invention discloses a CNN-LSTM-based magnetotelluric signal noise suppression method and system. The method comprises the steps of constructing a noise sample library and a pure signal sample library of magnetotelluric signals; training a convolutional neural network (CNN) by using the sample signal to obtain a CNN signal-to-noise identification model; inputting the actually measured magnetotelluric signal into a CNN signal-noise identification model to identify an interference data segment and a non-interference data segment; training a long-short term memory neural network (LSTM) by using the non-interference data segment to obtain an LSTM prediction model; selecting a non-interference data segment adjacent to the interference data segment, and inputting the non-interference data segment to the LSTM prediction model for loop prediction to obtain prediction data; and finally, carrying out data reconstruction on the prediction data and the non-interference data segment to obtain a denoised magnetotelluric