Background Estimation Based on Fuzzy Model for Thermal Video Surveillance

Background extraction is considered the first and key processing stage in video surveillance systems. This paper proposed an novel fuzzy architecture for background estimation. fuzzy architecture was built based on mamdani based fuzzy model and we used mean ,median and mode values as membership function. Background estimated by using this system is robust to adaptive background, and varying illumination conditions. Our experimental results demonstrate that proposed system is more efficient, robust and accurate than classical approaches.