Many fire detection systems are existed but itrequires some specialize hardware use of this hardware make system costlier so there is need of an efficient and inexpensive system. The main intent of propose system is to recognize fire through video surveillance camera to reduce ecological and economical losses. Existing systems narrates a method which is able to detect fires by analyzing the videos acquired by surveillance camera. Most of the them are having performance issues so these paper introduce some novel idea to overcome these performance issues and enrich the fire detection process. Proposed system describe main novelty about fire detection: complementary information respectively based on color analysis, shape variation and motion analysis are combined by multiexpert system. With the use of multiexpert system the detection of fire is done in the fast way as well as in accurate manner. System uses various algorithms by means of inputting the computed attributes and outputting the decision regarding the fire’s presence.

Keywords : Multiexpert system,fuzzy logic, Video surveillance.