Research of Fire Smoke Detection Algorithm Based on Video

Nan Zhao ； Xiao-Wei Wang ； Shou-Lin Yin

《International Journal of Electronics and Information Engineering》 13卷1期 (2021/03) Pp. 1-9

https://doi.org/10.6636/IJEIE.202103_13(1).01

摘要

Fire is one of the major disasters that seriously endanger the safety of human life and the natural ecological environment. The timely warning of fire is of great significance to reduce the loss. This paper presents a fire smoke detection algorithm based on improved background dynamic update and dark channel prior. It solves the cavity problem caused by the slow diffusion of smoke in the traditional motion detection algorithm and can eliminate most of the interference targets in the extraction stage of the candidate smoke area. Finally, the linear fusion of smoke color feature, rotating invariable LBP texture feature and HOG feature was used to identify by the nearest neighbor classifier (KNN). Experiments in multiple video scenes showed that the algorithm was less affected by environmental factors and had good smoke detection capability.

關鍵字

Background Dynamic Update ； Candidate Smoke Area ； Dark Channel Prior ； KNN Classifier