Emergency Evacuation Simulation Study Based on Improved YOLOv5s and Anylogic

With the development of social economy and continuous growth population, emergencies within field stations are becoming more frequent. To improve efficiency emergency evacuation further protect people’s lives, this paper proposes a method based on improved YOLOv5s target detection Anylogic simulation. This applies network to problem for first time, using stronger capability solve unstable data collection under unexpected conditions. uses YOLOv5s, which incorporates SE attention mechanism, detect pedestrians inside site. Considering height camera inability capture whole body pedestrian when site is crowded, adopts pedestrian’s head determine specific location ensure that task completed in shortest possible principle closest distance evacuation, so each can leave through exit him or her. The experimental results show average accuracy model incorporating mechanism reach 94.01%; constructed quickly provide an plan guide from nearest emergency, effectively verifying feasibility method. be extended applied research related construction aid decision-making systems stations.