Design And Implementation Of Fire Protection System Based On Multi-source Heterogeneous Perception Data Fusion

Fire is a great threat to human life and property safety.Therefore,the prevention and early detection of fires is very important to the safety of human life and property,especially for schools,homes and parks and other places where combustible materials are intensive.Nowadays,various smart fire protection systems have emerged,involving video surveillance,flammable gas leak detection,smoke detection,and so on.Some fire-fighting systems mainly rely on various sensor devices to monitor various physical quantities in relevant places,and some fire-fighting systems monitor based on video image analysis.Today’s fire protection systems are mainly faced with the following main problems:First,the demand for sensor equipment in the fire protection system is large,the sensor data format is not uniform,the second is that the monitoring is not comprehensive and accurate,and the last is the high manual consumption and intelligent not tall.To this end,this paper will carry out related research and work in the following areas:First,The sensor data format is not uniform.In this paper,a solution for sensor data configuration analysis is proposed.Users can customize the data packet analysis script.According to the received data packet,the system selects the corresponding script to parse the data packet,and the data Unified upload to the upper layer.Secondly,The demand for sensor equipment is large,and fire monitoring is not comprehensive and accurate.In this paper,a comprehensive analysis of sensor equipment monitoring data and video image data is proposed.Cameras and sensor equipment complement each other and can be reduced to a certain extent.The demand for sensor equipment in the fire protection system makes fire monitoring more comprehensive and accurate.Finally,it consumes a lot of labor and is not intelligent.This paper introduces the use of deep convolutional neural network architecture to automatically detect fire pixel areas in real-time video or static images,which is more intelligent and saves manpower.In this paper,a fire protection system is designed and developed.The system uses data fusion methods to comprehensively analyze various sensor data and video image data,so that fire detection is more comprehensive,accurate and intelligent.The related work results of the thesis have been effectively verified in the postgraduate innovation and entrepreneurship project of the school.