Research On Fire Detection Fusion Algorithm Based On City Shopping Center

In modern society,with the rapid development of economy and continuous update of science and technology,there are more and more large buildings,concentrated areas of commercial and residential buildings and high-rise buildings,with larger and larger scale and more and more complex architectural design and internal structure.Due to the particularity of these buildings,the degree of fire hazard and the difficulty of fire control and rescue are also increasing.Under such a development background,it is urgent to develop a fire detection system with higher performance and intelligence and sensitivity.Due to the unpredictable and non-structural characteristics of fire information,the false alarm rate cannot be reduced due to the large number of interference factors in the traditional singleparameter detector detection method.In recent years,many technical improvements have been made in the reliability and sensitivity of fire detectors to improve the accuracy of fire alarm to a certain extent.However,the practice shows that the current fire detection system has a lot of scope for improvement in automation and intelligence.To reduce fire false alarm,an important and effective research direction is to obtain more accurate and detailed description of fire characteristic information and more comprehensive and intelligent information analysis and processing when judging fire situation.Nowadays,many scholars and engineering practitioners are studying the fire detection technology based on multiple criteria and the intelligent methods it applies.Based on the application background of shopping center with its fire characteristics,this paper proposes an automatic fire detection alarm system based on data fusion algorithm based on the study of fire principle and current fire detection methods.Based on the characteristics of detection information and field environment,the system output the optimal decision automatically.For judgment and decision phase of fire detection algorithm,retained the common criteria(different types of fire probability)as one of the judgment reference,the real-time relevant background information of the fire detection area is also involved in the decision reference range of the fire decision-making,through various effective integration of information,expanding the decision factors category of fire detection,make the system decision-making level to a new level.The main contents and conclusions of this paper are as follows:Analyze the formation?development?mathematical model of fire,fire detection principle and major fire detection algorithm.Analyze the fire characteristics of shopping center in the city.Introduce the basic theory?process?structure and processing methods of different levels of data fusion.Analyze the advantages of this technique used in the field of fire detection.Establish a multi-source information system framework model of fire detection fusion algorithm,from different levels of fusion processing stage?degree of information abstraction,fire detection fusion algorithm is divided into three different levels(information layer fusion algorithm,feature layer fusion algorithm,policymakers layer fusion algorithm),the specific information fusion processing steps and methods are different.The main tasks of the information layer: fire detector through the first set various types of sensors to collect appropriate and related fire characteristic parameters(commonly used with temperature,smoke,CO),and then finish the pretreatment to the original data,using local decisions to distributed processing of information,once the characteristic parameter display abnormal,the same set of features parameters will be submitted to features layer for further identificationThe main task of the feature layer: using the BP neural network algorithm based on l-m,the physical parameters input into feature layer are fused and identified,and the probabilities of various fires(probability of open fire?smoldering fire)judged at the current moment are output.The main task of the decision-making layer: using the fuzzy reasoning algorithm,give full consideration to the current regional environment conditions related to fire relevant information,introduce the indirect criterions(fire hazard and fire risk)assessment,and make the fire detection system can combine the environmental conditions of the detection area(area assistant decision factors),with the direct criterion output by feature layer and duration to fuse and process,and the way people to experience the fire determination?consider expressed in the form of fuzzy control rules,propose 162 fuzzy inference control rules,through fuzzy logic reasoning,automatically output system decision with higher reliability.Through the simulation of MATLAB experimental fire data,it is proved that the method proposed in this paper can make the feature layer of the fire detection system have good recognition ability and the decisionmaking level of the decision-making layer can be effectively improved,and is also applicable to the fire detection system in the shopping center of the city.