CNN-Transformer for Microseismic Signal Classification

The microseismic signals of coal and rock fractures collected by underground sensors contain masses blasting vibration generated mine blasting, the waveforms two are highly similar. In order to identify true with a monitoring system quickly accurately, this paper proposes lightweight network model that combines convolutional neural (CNN) transformer, named CCViT. Of these, CNN is used extract shallow features locally, transformer deep globally. Moreover, modified channel attention module provides important information for suppresses useless information. experimental results on dataset in show proposed CCViT has significant advantages floating point operations (FLOPs), parameter quantity, accuracy compared many advanced models.