The Fire Detection System Based On Video Monitoring Research And Implementation

The dangers of fire are obvious.With the development of the society and the proposed of The Peaceful city, the fire detection is also paid great attention. The requirements for automatic fire detection system become more stringent. The traditional fire detector already cannot satisfy the development of the society. But for the security industry, all kinds of video monitoring system has spread to every corner of the city. So it will be at a lower cost to improve the fire detection ability of the city, if we can use the security monitoring system to realize fire detection. Therefore, Using the security monitoring system to realize the fire detection, and realize a monitoring network which combine the security monitoring with fire detection. It has important practical significance for the intelligent and networked development of fire detectionAfter analyzing the imaging features of flame and smoke under the visible light and infrared system, the paper design a fire detection system based on video monitoring. This system mainly divided into three modules, including monitoring module, detection module and alarm module. Monitoring module is mainly responsible for security monitoring. Detection module and alarm module are the heart of the system, it is mainly responsible for the fire detection and fire alarm of video fire detection system. When the security monitoring is work,at the same time implement the detection of the flame and smoke under visible light, if the result of fire pre-determination is fire, tracking the target, and then immediately start the infrared filter of the camera to detect the infrared temperature detection. Using two stage judgment model can reduce the rate of false positives of the system.when the security monitoring is work,at the same time implement the detection of the flame and smoke under visible light, if the result of fire pre-determination is fire, then immediately start the infrared filter of the camera to detect the infrared temperature image, when the result meets the temperature characteristics, then immediately start the fire alarm mode. Using two levels of fire determination model can reduce the false rate of the system.In the algorithm research, the paper has proposed a algorithm about moving object detection based on four frame difference method and mixture gaussian background mode. Using the wavelet transform to draw the high frequency energy and correlation of the smoke area, and Proposing a algorithm about smoke detection using the high frequency energy and correlation between video frames. Finally, according to the comprehensive detection of dynamic and static characteristics of the smoke to achieve the goal of fire smoke detection.For the flame detection, using Comprehensive detection of dynamic characteristics,static characteristics and the infrared temperature characteristic to realize the detection of flame.The main characteristic of this system is combined security monitoring and fire detection. Using two levels fire determination model,which use visible light to pre-judgement and Infrared temperature detection to judge again.Acoording to the detect result of two levels of fire determination model to start the fire alarm module. Using the VS2010 and opencv2.4.2 to implement the fire detection system under the Win 7 operating system, and Collect related video test the system.