Deep goaf CSAMT electrical characteristic enhancement and classification method based on machine learning

The invention discloses a deep goaf CSAMT electrical characteristic enhancement and classification method based on machine learning, and the method comprises the steps: I, initializing a convolution kernel which can remove shallow abnormality and enhance the characteristics of a deep goaf according to the resistivity abnormality characteristics of CSAMT after inversion; II, carrying out convolution calculation by utilizing the convolution kernel obtained in the step I so as to extract shallow abnormal information and deep goaf electrical abnormal information; III, calculating the shallow convolution resistivity characteristic and the deep goaf convolution resistivity characteristic obtained in the step II to obtain an energy error, and if the energy error does not meet the precision requirement, returning to the step I to modify and initialize convolution kernel parameters; if the energy error meets the error requirement, removing shallow abnormal interference, and enhancing the electrical characteristics of