Open    Access



Fire Detection in Video Sequences Based on Static and Dynamic Features



Author(s)

Rykhard Bohush, Nadeja Brouka, Sergey Ablameyko

Full-Text PDF XML     288 Views

DOI:10.17265/2328-2223/2013.12.004

In this paper we propose the efficient smoke and flame detection algorithms for intelligent video surveillance systems. Our algorithms consider dynamic and static features of smoke and flame including contrast, color, texture features and motion information. For smoke detection, the approach uses the motion and contrast as the two key features of smoke. Motion is a primary sign and it is used at the beginning for candidate areas extraction from a current frame. Furthermore, the direction of smoke distribution is also considered as an estimated movement based on the optical flow technique. For flame detection, we use the color image segmentation technique. Firstly, the temporal and spatial wavelets are analyzed. Then, color and texture features for candidate flame regions are extracted. Texture features are defined based on the normalized gray level co-occurrence matrix after computation of local binary pattern. Experimental results asserted the advantages of the proposed fire detection in the diversity of test video sequences.

Flame and smoke detection, background subtraction, contrast analysis, wavelet transform, texture segmentation