Fire Detection in the Field of Color and Fire Motion, Using the Differential Method

Fire is one of the significant problems of today s world and creates irrecoverable damages for humanity so that providing a smart fire alarm and extinguish system is inevitable. Currently, there are many different fire alarm systems but they are not recommended due to limited mechanical function and/or life cycle as well as the high costs and limitation of detecting distance. As today security cameras are installed everywhere, using them as fire alarm system will not only reduces the costs significantly, but also removes the abovementioned disadvantages remarkably. This paper attempts to detect fire quickly with high accuracy. In order to do this, first, the color zone of fire in YCbCr and RGB color system which contributes to finding suitable candidate point for fire in these systems based on natural color is studied. Regarding the fact that studying color zone is faster than movement zone, this method, first, investigates color zone and will consider the movement zone, in case if there is a suitable candidate. Movement zone deals with fire movement and its differentiation in this paper. In this study, only those points are considered fire which have movement and the color characteristics of fire. In this method which has been performed in color and movement zone a relative improvement in fire detection is observed and also there is a relative increase in TP and FP values, compared with those in the previous papers. The average of improvement of this method is about 0.86%, compared with Tung method