An approach for the compensation of the background variability in fire detection systems

Abstract

Background (BG) variability is an important issue in fire detection. Due to atmospheric situations in general a high correlation can be assumed between the BG infrared radiation for spatially near zones. This correlation can be even greater if similar forest and soil-surface conditions exist in these zones. In this paper we present an approach for fire detection in ground systems. This approach takes into account the temporal and spatial BG behavior. We show that if we use differences of sensor outputs of BG zones with highly correlated surface temperature and reflectance then the false alarm probability is reduced for a given threshold. We also show that for a fixed false alarm probability the use of differences between sensors instead of a single sensor provides higher fire detection probability.