Title :
Classification and detection of fire on WSN using IMB400 multimedia sensor board
Author :
Pande, Vijae ; Elmannai, Wafa ; Elleithy, Khaled
Author_Institution :
Dept. of Comput. Sci. & Eng., Univ. of Bridgeport, Bridgeport, CT, USA
Abstract :
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.
Keywords :
alarm systems; computerised monitoring; emergency management; filtering theory; fires; forestry; image classification; image colour analysis; multimedia systems; visual databases; wireless sensor networks; IMB 400 multimedia sensor board; WSN; classifier training; color information; damage monitoring; early alarm system; early detection system; energy consumption; environment degradation; filtering algorithm; fire classification; forest fire detection systems; human resources; image database; natural damage; natural resources; property damage; Cameras; Fires; Image color analysis; Monitoring; Multimedia communication; Wireless sensor networks;
Conference_Titel :
Systems, Applications and Technology Conference (LISAT), 2013 IEEE Long Island
Conference_Location :
Farmingdale, NY
Print_ISBN :
978-1-4673-6244-3
DOI :
10.1109/LISAT.2013.6578247