A New Approach to Vision-based Fire and its Intensity Computation Using SPATIO-Temporal Features

Currently, fire detection systems based on computer vision techniques are highly appreciated for their intelligent detections at earliest. These systems use surveillance cameras to capture high-level information from a fire that enables a system to take preventive and corrective measures before the happening of a fire hazard. Handling false fire detection and reducing false alarm in such developed systems are still big challenges that need to be addressed. In this paper, a novel framework is proposed that uses angular and regional area information of the fire flame to predict the existence of a fire flame in sequence of a video frames. The proposed system especially handles false detection of fire object in video. The results achieved on different dataset of fire videos show that extracted features using the proposed framework efficiently distinguish fire and non-fire objects.r These features are also useful to estimate the size and direction of a fire flame. The regional area information can be fed into a machine-learning algorithm to learn a model for a fire flame that can be used to predict about the existence of fire flame.