الفهرس Only 14 pages are availabe for public view
 |  | from | 117 |  |  |
| | | from | 117 | | |
Abstract
This thesis introduces various methods of fire and smoke detection, based on different static and dynamic characteristics features of fire and smoke. Two fire detection algorithms were proposed. The first algorithm is based on the color, static feature of the fire in the visible range images, which consists of four sub-algorithms: (i) fire detection using CMY color space, (ii) fire detection using CIEL*a*b color space, (iii) fire detection using the yellow color static characteristics in RGB color space, (iv) fire detection using both RGB and HSV color spaces. These sub-algorithms output decisions are then connected to a decision fusion weighted voting system, that reaches the final decision of fire detection or not. The second algorithm is based on fire edge detection. Three smoke detection algorithms were proposed. The first algorithm is based on the color static feature of smoke in the visible range images, detecting low and high densities of smoke. The second algorithm is based on smoke edge detection. The third algorithm is based on wavelets decompositions and energy analysis of the edges of the images or the video frames, and it represents a study of the effects of smoke on images. The Experimental results show a study of their accuracy, detection rate and false detection rate