EARTHQUAKE EVENT CLASSIFICATION METHOD USING ATTENTION-BASED CONVOLUTIONAL NEURAL NETWORK, RECORDING MEDIUM AND DEVICE FOR PERFORMING THE METHOD

摘要

An earthquake event classification method using an attention-based neural network includes the steps of: centering and preprocessing input earthquake data; Extracting a feature map by performing nonlinear transformation of the preprocessed seismic data through a plurality of convolutional layers of at least three; Measuring the importance of the learned feature based on the attention technique modeling the interdependency between channels of the feature map with the nonlinearly transformed seismic data; Correcting the feature value by multiplying the measured importance value with the learned feature map for each element; Down-sampling through max-pooling based on the feature value; And classifying an earthquake event by normalizing the down-sampled feature value. Accordingly, through attention-based deep learning, key features inherent in a large amount of/complex data are extracted, and through this, it overcomes the limitations of the existing micro-earthquake detection technology and enables earthquake detection even in a low SNR environment.