Early Smoke Detection in Outdoor Space by Spatio-temporal Clustering using a Single Video Camera

The early smoke detection in outdoor space concerns to a security sphere and is realized in landscape monitoring systems, in video surveillance systems near building, on bridges, ships, into tunnels and etc. We have suggested a novel video-based method of smoke detection by spatio-temporal clustering which involves three developing stages. The first stage connects with motion detection, the second stage is based on a color-texture analysis of moving regions, and the third stage is enhanced by a spatio-temporal clustering of moving regions with a turbulence parameter. A spatio-temporal volume data permits effectively track a dynamics of smoke propagation in outdoor space by using the designed software functioning in a real-time mode. Experimental results show that the proposed set of spatial and temporal features well discriminate smoke and non-smoke objects in outdoor scenes with a complex background.