Vision based Fire/Flood Alarm Surveillance System via Robust Detection Strategy

This paper presents an intelligent and robust event detection method on visual surveillance system. Based on real time video content analysis, potential dangerous events such as fire and flood can be detected automatically in the early stage thus to prevent from damage. Moreover, to improve the correct detection rate, a robust strategy is included to reduce the false alarm rate due to effect of varying lighting condition and complex background illumination. Compared with the existing systems, the experimental results show that the proposed system not only has better performance but also alleviate the system cost, and is suitable for practical implementation.