When a disaster occurs,it is important to evacuate the crowd safely and efficiently.The effective and feasible evacuation simulation method has important theoretical and practical significance.Evacuation simulation technology models the behavior of the crowd in emergency situations,makes reasonable predictions of the evacuation behavior of the crowd,assists in the formulation of large-scale crowd evacuation programs,and reduces public safety hazards.The current mainstream methods are less concerned with the dynamic evolution of disasters related to evacuation and avoidance psychology of crowd.This dissertation mainly studies the dynamic evacuation simulation methods under the influence of disaster and avoidance psychology.The main work of this dissertation is as follows:1)The current situation and development trend of the crowd evacuation simulation are reviewed.The current mainstream evacuation simulation algorithm model and the algorithm for extracting evacuation path from image are analyzed and summarized.2)An improved method for constructing evacuation maps from aerial imagery of drones is proposed.In order to simulate evacuation based on road information in real environment,the D-LinkNet model is used to extract the walkable area in the aerial image,and the post-processing of the image is used to correct the connectivity of the path information,and the extracted evacuation path information is optimized.Then the corresponding evacuation map is generated based on the extracted evacuation path.3)A dynamic Evacuation Route Replanning Method(DERR)is proposed.Firstly,the above algorithm is used to generate the evacuation map from the aerial image;then,considering the dynamic evolution of the disaster related to evacuation and the avoidance psychology of crowd,the A* pathfinding algorithm is used to re-plan and optimize the evacuation path of the crowd.The evacuation simulation path is dynamically generated by continuously updating the disaster state and evacuaton map.The validity and authenticity of the proposed method are verified by experimental simulation of the evacuation process under the dynamic evolution of disasters and the analysis of psychological avoidance parameters. |