An Intelligent Algorithm for Determining Optimal Wildfire Suppression Zone Using UAV

Wildfires inflict profound environmental and economic devastation. Aerial firefighting is a critical component in wildfire suppression efforts, offering the ability to access remote areas and take quick action. Traditional aerial firefighting operations totally depend on the experience and subjective judgment of pilots to determine the wildfire suppression zone and drop the fire retardant on it, resulting in inconsistent performance and a lack of scientific guidance. With the optimization objectives that cover as many wildfire points as possible within a limited zone, based on wildfire point location information observed by UAV and aircraft's motion state and related parameters, a Rectangular RANdom SAmple Consensus and Ground Pattern (R-RANSAC_GP) algorithm combining R-RANSAC and the Ground Pattern method is proposed to determine the optimal wildfire suppression zone in real time. Both the simulated wildfire point and the real wildfire images are used to test the performance of the R-RANSAC_GP. Results demonstrate that R-RANSAC can determine the optimal retardant dropping zone autonomously and intelligently.