Forest fire detection by using color feature of fire/ flame

Forest fires are a disaster that affects the atmosphere in the negative way, destroying the habitat of many living species and destroying plant and animal species. In the fight against forest fires, the early detection of fire and the introduction of warning systems have a great important. Nowadays, depending on the development of the technology, image processing techniques are used in fire / flame detection. Usually, static or dynamic characteristics of fire/flame are used in developed methods for fire / flame detection. In this study, the color features of the fire images were used and the fire images transformed to a new color space to detect the flame. The obtained new color space is developed by using the Artificial Bee Colony (ABC) algorithm. Otsu thresholding method was used on converted images in new color space. Flame detection was performed on binary images. After the flame detection process is completed, the obtained binary images are subjected to the similarity measure with the real binary images contained in the data set and the success criterion of the proposed method is calculated.