Research On Embedded Fire Image Detecting And Auto Fire Extinguishing System

Fire is one of the highest frequency, the heaviest destroy and the strongest influence disasters in modern city. For quick alarm and auto fire-fighting, timely controlling and extinguishing, and minimizing casualties and property losses, equipping fire alarm and automatic fire extinguishing system has become necessary in the industrial and civil buildings, shopping malls, hotels, libraries, scientific research and commercial sectors. Fire alarm and automatic fire extinguishing system in modern building plays an extremely important role in the safety and security.Fire image detection and fire extinguishing system integrates a lot of high-tech, such as fire science, computer vision, image processing technology, pattern recognition and artificial intelligence, it is highly efficient with the ability to control fire. Now, these are mainly products with fire image detection and automatic fire extinguishing mutual separation, independent layout.Some integrated products in the special environment are expensive, complex, and not suitable for large-scale promotion. On the basis of research on fire image detection, recognition, and fire locating and controlling methods, this paper organically integrates fire image detection and automatic fire extinguishing, by applying ARM Embedded technology. It simplifies the layout of fire fighting equipment and reduces costs.The paper gives an overview of the current domestic and foreign R & D status about fire detection and fire fighting technology, and researches on fire image pre-processing technique, such as video image enhancement, filtering and morphological filtering etc. At first, gain the information of surveillance scenes by traditional background subtract method, and then real-time updates on the background by improved Background update algorithms. A new fire image segmentation method which bases on GA optimization of Otsu is applied in the project, it can effectively segment fire flame image.Through analysing the characteristics of fire flame image and extracting the characteristics such as fire growth and spread, similarity and the edge of jitter, give the fire recognition algorithm of the fire, and carry on the corresponding experimental verification.After achieving the Fire recognition, automatic fire extinguishing ought to be carry out. Accurately achieving spatial location of the fire source is the key technology to automatic fire extinguishing. Also after establishing the geometric model of the camera, get the mapping between the image coordinates and spatial coordinates, research on the camera calibration methods, and give specific calibration algorithms. According to the principle of binocular tracking and on the basis of the method of monocular space positioning, a new method of using geometric lines to calculate the three-dimensional coordinates is proposed. Thereby, enhance system adaptability and improve positioning accuracy.On the basis of researching on the fire image detection technology and the automatic positioning fire-fighting methods, this paper design overall project of system in accordance with the feature of early fire and the technical requirements of system. According to the functions, system is divided into several modules. Discuss in detail the design of the core ARM embedded chips (S3C2410A) and the various modules hardware and schematic design. Complete the system's hardware design and system's software modular design, and realize the modules flowchart. Finally, on the designed platform of system, perform experiment about fire recognition and space location, test the feasibility and reliability of the system.