Research On Multi-information Fusion Model And Algorithm For Smart Fire Fighting

Now,fires have gradually become one of the main accidents in human life,and the occurrence of fires has also brought huge cost to the society.Therefore,fire protection has become an important attribute to measure urban safety.Fire protection involves multiple systems,multiple departments,and multiple links.However,the current islanding of fire management information can easily cause safety hazards.To ease the deficiency of traditional fire detection,the concept of "smart firefighting" is raised.For this requirement,thesis considers the two perspectives of "elimination" and "prevention",and studies the multi-information fusion model and algorithm for smart firefighting.First of all,the "elimination" of fire is intended to eliminate fire,and early detection of fire can make the damage of fire less and easier to eliminate.Single signal detection of fire is susceptible to other factors,causing false alarms.Therefore,thesis introduces a new fire detection data source and uses many source fusion to propose a fire detection model.The model combines the Jousselme distance to assign weights to the data sources,and finally uses D-S evidence theory to fuse multiple discriminant information.The consequence reveals that,in the identical factors,thesis’ s model detects the occurrence of fire 3-5 seconds earlier than the present method,fire detecting time is advanced.Secondly,build a fire detection demonstration system.The system signal acquisition device uses the existing development board,visualization platform based on Spring、Spring MVC and Mybatis frame structures,uses the Java language transplanting research thesis’ s model.The system can acquire and transmit signal,detect fire and warn in time.Finally,for the prevention of fire,thesis set out from risk assessment.Considering that old community don’t have enough fire control device,the happen rate and harm of fire are bigger than other scenes.So,thesis study the old community.Thesis build the fire risk estimate target system of old residential quarters from two perspectives of "static" and "dynamic".On this basis,this thesis use AHP and entropy method to term target weight,and use fuzzy method to evaluate the possibility of fire accident in old community.And choose a community of Shanghai as an example to verify the practicability of the model.