B008 Fire smoke image entropy

Smoke and fire image segmentation using Shannon’s entropy method and mathematical morphology

Brief:

In order to improve the control of smoke management system in large space, new visual smoke and fire detection techniques is proposed. Automatic segmentation of smoke and flame image from the background images is an important element in visual smoke and fire detection. For most visual images obtained for fire and smoke, the dual peak obtained from the histogram is not significant and unable to carry out a correct segmentation. In the present study, Shannon’s entropy method and mathematical morphology are used for segmentation of smoke and flame images. This study showed Shannon’s entropy and mathematical morphology can be used to extract the information related to the flame accurately form the experimental result. This formed the basis for visual smoke and fire detection and would be useful for fire analysis.

Further information:

Li J, Fong NK, Wong LT, 2003. Smoke and fire image segmentation using Shannon’s entropy method and mathematical morphology, Proceedings of Shandong-Hong Kong Joint Symposium 2003, Qingdao, China, The Hong Kong Institution of Engineers, 17-18 October, pp. A198-A207.