Every year, accidental fires cause numerous properties damages as well as human losses; and this situation has been continuing for decades. Even though in today’s modern world, innovations and new technologies are already applied to early detect smoke or fire; there are some limits to these solutions, from which we could recognize potential areas to improve. In this paper, FPT’s Analytics team proposes a solution to early detect both fire and smoke using video sequences captured by a single camera. The solution is based on the static and dynamic features of fire/smoke in the temporal and spatial domains as color feature, slow motion and growth to recognize and detect fire/smoke in the outdoor environment. The result of this research can help fire departments and rangers improve the efficiency of fire monitoring & management, leveraging the use of Artificial Intelligence.
Download Full Version of the White Paper