Unsafe Mining Behavior Identification Method Based on an Improved ST-GCN

Aiming to solve the problems of large environmental interference and complex types personnel behavior that are difficult identify in current identification unsafe mining areas, an improved spatial temporal graph convolutional network (ST-GCN) for miners’ a transportation roadway (NP-AGCN) was proposed. First, skeleton spatial-temporal map constructed using multi-frame human key points used recognition reduce caused by environment coal mine. Second, aiming problem original structure cannot learn association relationship between non-naturally connected nodes, which leads low rate climbing belts, fighting other behaviors, reconstructed partitioning strategy changed improve ability model multi-joint interaction behaviors. Finally, order alleviate convolution has difficulty learning global information due small receptive field, multiple self-attention mechanisms were introduced into In verify detection regarding identifying behaviors mine belt area, our tested on public datasets NTU-RGB + D self-built area. The accuracies proposed above 94.7% 94.1%, respectively, 6.4% 7.4% higher than model, verified had excellent accuracies.