Abstract: In order to quickly and effectively monitor forest fires and improve the efficiency of forest fire detection, this paper proposes a technique for quickly identifying initial fire onset in forests by using unmanned aerial vehicles equipped with gas detectors. The technique uses unmanned aerial vehicles equipped with gas detection equipment for real-time monitoring of gas components and solid particle concentrations produced when a forest fire occurs in a simulated forest fire site, and studies the correlation between the variation of gas components and solid particle concentrations and the occurrence of forest fires. The research data showed that during the initial stage of forest fire combustion, the content of NO₂ and O₃+NO₂ gases and three types of solid particles (PM1.0, PM2.5, PM10) would rapidly increase, peak during the mid-combustion stage, and then gradually return to normal ranges post-combustion or after extinguishment. The content of gases and solid particles decreased as the distance from the monitoring sites to the fire source increased. This experiment can provide useful data for accurate monitoring of forest fires by observing the average gas content and trend of solid particle concentrations at various heights, thereby enabling the qualitative measurement of emissions during a forest fire.

Key words: unmanned aerial vehicles, forest fires, forest fire monitoring, gas survey, early warning