FireAnalyst: An effective system for detecting fire geolocation and fire behavior in forests using mathematical modeling

Authors: YÜCEL GÜLLÜCE, RAHMİ NURHAN ÇELİK

Abstract: This study proposes an effective new fire detection method and monitoring software for an early-warning fire detection system aimed at valuable forested areas, such as botanical parks or high conservation value forests, particularly those with boundaries. These critical forested areas need to be appropriately managed because they contain large concentrations of biological diversity, including threatened or endangered species, which are very susceptible to fire outbreaks; thus, early detection of fire and rapid response have a very important place in the fight against fire in those areas. In this proposed system, special detectors with state-of-the-art multispectral infrared technology and mathematical modeling algorithms have been utilized to create a smart fire detection system that can detect fires at a very early stage. The geolocation and behavior of emerging fires in a forest are also estimated with maximum spatial resolution by superimposing the detection areas of multispectral infrared detectors. In this study, candidate fire regions are examined for feasibility first. Next, the most suitable fire detector type is determined and used for expanding the fire control area, to have the highest positional accuracy in estimating the location of an emerged fire. Thereafter, mathematical models for the positioning of the detectors are created to have high spatial resolution in detecting the coordinates of forest fires by using the libraries of Google Maps APIs in the cloud. The geolocation of the fire and behavior of the fire inside the model are then simulated visually on the map portal, thanks to an extraordinary standalone software program called FireAnalyst. The proposed system was implemented for the Faruk Yalçın Zoo and Botanical Park in Darıca, Turkey. Experimental results have indicated that monitoring fires with FireAnalyst using selected multispectral infrared detectors positioned toward the center geometry outperformed other fire monitoring systems, providing a significantly shortened fire detection timeframe and high spatial resolution (up to 4.5 m) in detecting the geolocation of a fire in a minimum of ~3599.56 m2 forested area, and it adds functionality, such as real-time fire behavior analysis (spreading speed of fire, spreading direction).

Keywords: Wildland fire, fire geolocation, fire behavior, high conservation value forests, SMART method, GIS

Full Text: PDF