A Study on Smoke Detection Based on Multi-feature

Smoke detection based on video had great signification for early fire-incident alarm. In order to improve the accuracy of smoke detection, this paper proposed a smoke detection method based on multi-feature classification. Firstly, the smoke regions were segmented by three-frame difference using a temporal sliding window method. This method could overcome the shortcomings of the traditional algorithm in thin smoke region detection. Next, based on the obtained smoke features over frames, the real smoke regions were detected by Mahalanobis distance measure considering the dynamic property of smoke change. The experimental results show that the proposed method can detect the smoke effectively in the complex environment, and have good effect on the thin smoke detection.