1. A new strategy for fire detection based on analyzing the irregularity pattern of image contour.
- Author
-
Oghabi, Emad, Moghaddam, Reihaneh Kardehi, and Kobravi, Hamidreza
- Abstract
Fire is one of the most uncontrollable phenomena and is highly likely to result in irreparable harm if it is not detected at its early stages. A lot of researches have been conducted to propose more efficient fire detection methods. In this paper, in order to facilitate fire detection and propose a more cost-efficient and also a highly accurate method, background color of images based on RGB and YCbCr color spaces is examined. In this regard, some suitable candidate regions for fire are chosen. Then, changes of boundary irregularities of these selected contours are measured by calculating the largest Lyapunov exponent. These irregularities of selected images can be considered as an important factor in early fire detection. If the measured irregularities are greater than 1000 fire videos, the selected candidate will be considered as an image of fire. Simulation results prove that the proposed algorithm can lead to more accurate detection of the fire region with acceptable false detection rates. [ABSTRACT FROM AUTHOR]
- Published
- 2024
- Full Text
- View/download PDF