Approach To Abnormal Behavior Recognition And Escape Route Planning In The Emergencies Of Public Buildings

The urban population in China is increasing along with the continuous improvement of urbanization rate.Large public buildings in the city are getting more and more in order to meet the living needs of people.However,in recent years,the frequent occurrence of public building emergencies causes life and property loss and adverse social impact.Therefore,the emergency management of public buildings emergencies has gained great attention.The present intelligent video surveillance technology is mature and has been widely used in security and other fields.However,there are still many problems in the practical application of abnormal behavior recognition.Most of the researches on path planning focus on the optimization of algorithms and the path planning of robots,and the application of escape route planning in public buildings also needs to be explored.In this paper,the abnormal behavior identification and escape path planning were realized by using the intelligent video surveillance technology and path planning technology,which are the basis for establishing a systematic,comprehensive and efficient emergency management platform.In this paper,we firstly analyzed the relationship between the public building emergencies and people's abnormal behavior through the actual public building emergencies,thus identified the key abnormal behavior that need to be monitored in the public buildings.Then,the motion,size and texture features of the target are extracted based on the Lucas-Kanade optical flow method and the 2D Gabor filter,we modeled the target subject motion and size using Gaussian kernel density estimation algorithm,while we modeled the target subject texture using the codebook algorithm.Then the abnormal behavior,e.g.,moving speed change,gravity center change and population density change,can be identified through the abnormal judgments in public buildings and issue early warning.Secondly,based on the A* algorithm,we proposed a path planning method based on public building grid partition.This method requires two conditions,namely both the total escape distance of the building's internal personnel and the difference among the number of the escaped people from each exit are as small as possible.Finally,we proposed three subsystems of the emergency management of public building emergencies,e.g.,early warning system,escape route planning system and information release system.And based on this,the basic framework of emergency management of public building emergencies is constructed.Experiments demonstrate that the proposed automated approach to abnormal behavior recognition and escape route planning for the public building emergencies have a certain impetus to the construction of early warning system and escape route planning system in emergency management of public buildings.The designed basic framework for the emergency management of the public building emergencies also provides theoretical support for the future emergency management platform construction in order to improve the efficiency of public building emergency management and reduce the loss of living and property caused by emergencies.