Vision Based Fire Detection

Abstract

Automated fire detection is an important research issue due to its significance in the protection of life, property and environment. Reliable detection and early warning is crucial for timely evacuation of people and goods to safety during fire emergencies. Indoor fire detectors use smoke and heat sensors however such detectors are not that effective in outdoor fire detection. With the development of intelligent video surveillance systems, a relatively newer technology of video fire detection has attracted considerable attention. The basic idea is to use image frames of a scene captured by a CCTV camera to determine the existence of fire. Various image processing and machine learning algorithms are applied to process the captured image frames in order to determine the possible existence of fires within the vision field of the cameras. Distinguishing features like color, motion and shape are used to detect fire and determine the fire and non-fire regions. The proposed system can be easily incorporated into building video surveillance systems and intelligent building systems.