Video fire detection makes a significant contribution to the effectiveness of fire detection systems, particularly as regards fire in large spaces such as Atria, Tunnels, Hangers, Warehouses and E&M Plant rooms, as traditional fire detection systems have been shown to be ineffective in large spaces. For the development of video fire detection systems, spatial, spectral and temporal indicators are important in the identification of a fire source. In the development of video fire detection systems, flame image segmentation, recognition, tracking and predication are important areas of investigation. The multi - threshold algorithm of Otsu's method and the Rayleigh distribution analysis method (modified segmentation algorithm) can be used in the segmentation of flame images. The modified segmentation algorithm, however, can be strengthen to extract the pool fire images making use of the optimum threshold values. Following such segmentation the pool fire images centroid analysis technique can be used to recognize pool fire images by means of the Nearest Neighbor (NN) algorithm. The objective of this paper is to examine the modified segmentation and the NN algorithms.
|Number of pages||12|
|Publication status||Published - 1 Jan 2014|
|Event||2013 International Conference on Performance-Based Fire and Fire Protection Engineering, ICPFFPE 2013 - Wuhan, China|
Duration: 16 Nov 2013 → 17 Nov 2013
- Video fire detection
ASJC Scopus subject areas