REAL TIME IDENTIFICATION AND LOCATION OF FOREST FIRE HOTSPOTS FROM GEO-REFERENCED THERMAL IMAGES

More than forty percent of Canada's land area is covered by forest, of which fifty-six percent is commercially productive. However, this forest resource is constantly at risk of destruction by wild-land fire. Fire annually burns 433,000 hectares of commercial forest, and it enlarges Canada's forest management budget. This paper briefly explains the application requirements and system components used to develop a system to potentially improve the response time of the forestry department in the identification of newly ignited forest fires. This project is currently under development by the mobile multi-sensor research group at the University of Calgary. The objective of this research is to develop a real-time Forest Fire Detection (FD) system for identifying and locating forest fire hotspots automatically using geo-referenced thermal images, and contribute to reducing the impact of forest fires through earlier detection. The system consists of an integrated WADGPS/INS unit for direct georeferencing purposes, and a thermal imager to detect potential forest fire hot spots. In this paper, the process to properly identify forest fires from thermal images is briefly explained, followed by the corresponding georeferencing results from using this process.