Classification and detection of fire on WSN using IMB400 multimedia sensor board

Over the years fire detection systems have been developed using multiple techniques. These systems monitor the damage done by forest fires and tend to reduce the environment degradation and save the natural as well as human resources. On the other hand, these techniques or methodologies still need a lot of effort because they are mostly a high cost maintenance process for early detection. Otherwise some systems have been considered as slow systems in detecting fires. It is a well documented fact that detecting a fire would not be enough for real time cases. An early detection and an early alarm system can rapidly improve the detection process and avoid loss of life as well as property or natural damage. In this paper we introduce a new detection system for fire detection using a multimedia board in order to detect and verify the fire in less time. The idea of the new algorithm is to add the capability of multimedia in an efficient way. Therefore, we used the IMB 400 Multimedia board in order to capture the images and run our filtering algorithm over the images to detect the fire. Hence, with the IMB 400 board´s sleep/wake up ability, we can save on the critical issue of energy consumption. With this system, we will be able to detect and verify the fire in an environment at the same time and save the fire images in a database for further training of the classifier. This would be a robust system itself. Lastly, in this paper we are showing the importance of color information and movement over the detection of fire systems by using the Multimedia board (IMB400) in our implementation.