Development of a Method for Extracting Coefficients of Spectrograms of Electrofacies based on MFCC

Abstract

This paper describes the method of preparing spectral data of electric logging for convolutional neural network models based on mel-frequency cepstral coefficients methodology as a part of automatisation of facies analysis. As an example, the allocation of coefficients for two types of facies is considered. The result of the research show the possibility of using the considered method for preparing spectral data of electric logging.