Miners’ safety awareness is crucial to mine safety production. Insufficient safety awareness among miners can easily lead to safety accidents. This study applies the DEMATEL-ISM model to explore the relationships among the multi-dimensional influencing factors of miners’ safety awareness. First, a causal matrix and network relationship influence diagram for each dimension and indicator are constructed to identify 12 multi-dimensional influencing factors of miners’ safety awareness and determine the key influencing factors. Second, based on a one-dimensional convolutional neural network, the SE-ResNetV2 model with a channel attention mechanism was constructed to achieve adaptive assessment of miners’ safety vigilance. The DEMATEL-ISM model clearly reveals the hierarchical relationships among the factors influencing miners’ safety vigilance, while the SE-ResNetV2 model reduces prediction errors to a certain extent. The relative prediction error of the model when inputting multi-dimensional data features is only 9.5%. This study provides new theoretical and technical pathways for ensuring safety in mine operations and has significant practical application value.
