Magnetotelluric signal identification and reconstruction method and system based on deep residual network

The invention discloses a magnetotelluric signal identification and reconstruction method and system based on a deep residual network, and the method comprises the steps: constructing a sample library 1, and marking the class label of each data segment; performing Graman angle field transformation on the data segments in the sample library 1 to obtain a Graman angle field image; constructing a deep residual error classification network, and performing network training by using the Graman angle field image and the corresponding classification label to obtain a magnetotelluric data classification model; segmenting to-be-denoised magnetotelluric data, then performing Graman angle field transformation, and inputting the transformed magnetotelluric data into the magnetotelluric data classification model to obtain a classification label of each data segment; carrying out reconstruction or denoising on the identified noisy data segments; and finally, splicing the identified high-quality data segments with the recons