Naslov
Forest Fires Detection in Digital Images Based on Color Features

Autori
Tuba, Viktor ; Capor-Hrosik, Romana ; Tuba, Eva

Izvornik
International Journal of Environmental Science (2367-8941) 2 (2017); 66-70

Vrsta, podvrsta i kategorija rada
Radovi u časopisima, članak, znanstveni

Ključne riječi
Image processing, YCbCr, fire detection, color features

Sažetak
: One of the biggest ecological catastrophes are forest fires. Early detections of the fires can have significant impact on the control of forest fires which is very important for successful extinguishing. Areal or satellite surveillance is one of the systems used for monitoring forest areas. Based on the images provided by this system, fire can be detected by using some image processing techniques. In this paper we proposed an algorithm for forest fires detection based on color features. Different characteristics of the color components of YCbCr color model were used to detect fire based on the predefined threshold values and combination of the components values. After pixel classification morphological operations were performed to remove incorrect classified pixels. Our proposed method successfully detected fire regions in forest images with different lightning conditions.

Izvorni jezik
Engleski

Znanstvena područja
Matematika, Računarstvo

POVEZANOST RADA