A new hybrid algorithm for fire vision recognition

This paper proposes a novel method to detect fire and/or smoke in real-time by processing the video data generated by an ordinary camera monitoring a scene. The objective of this work is recognizing and modeling fire shape evolution in stochastic visual phenomenon. It focuses on detection of fire in image sequences by applying a new hybrid algorithm that depends on optimizing the back-propagation algorithm, after canny edge detection, for determining the smoke and fire boundaries. Another clue is used in the fire detection algorithm that detects smoke and fire flicker by analyzing the video in the wavelet domain. Color variations in flame regions are detected by computing the spatial wavelet transform of moving fire-colored regions. Experimental results show that the proposed algorithm is very successful in detecting fire and/or smoke.