Research On Fire Detection System Based On Multilevel Hybrid Sensor

In recent years, with the continuous development of society, the living standards of people’s material and spiritual culture level are raised, and the use of fire is increased, but the fire occurrence frequency, harm and loss are increased. Therefore, the demand of fire detection and control system is greatly increased. Fire process can be divided into three periods, characteristics of the fire in the early stage is that small fire, low temperature, relatively easy to put out the fire and resulting in small loss and harm. In the middle of the fire, the fire is developing rapidly and the temperature rises rapidly, which is the main cause of fire damage and harm. In the final stage of the fire, the combustible was burnt, the temperature gradually reduced, and the loss is irreparable. Therefore, the key period of fire fighting is the initial stage of fire. It’s most important to detect fire and quickly send alarms and douse the fire timely.Firstly, the paper analyzed the occurrence process of fire. According to the physical characteristics and image characteristics in the process of fire, the fire detection scheme was determined. The first level detection system used an ionic smoke sensor and pyroelectric infrared sensor, respectively based on the smoke characteristics and spectral characteristics of the fire. A detection warning signal was transmitted to the camera sensor based on image processing, to further confirm the fire. If confirmed that the fire occurred, a warning message was send to the relevant person through the three level system of GSM SMS module. According to the design scheme of multilevel fire detection system based on hybrid sensor, the software and hardware design was carried out. The hardware design included:first level detection module design, power module design and remote alarm system design; Software design includes:first level detection module program design, serial programming, video capture and continuous frame access, short message alarm. According to the design requirements, the paper did the fire image processing research, including: image processing basic technology, suspected fire area identification, flame feature extraction. Finally, the fire image can be identified.In this paper, the hybrid sensing and multilevel detection method was applied to the fire detection system and warning messages were send through SMS module, which could make the fire detection information more accurate and timely, in accordance with the requirements of fire detection system.