DEEP LEARNING MODELS AND PERFORMANCE EVALUATION FOR FOREST FIRE DETECTION USING AERIAL IMAGERY DATA FROM UNMANNED AIRCRAFT
Abstract - Forest fires in Northern Thailand have been a persistent and significant issue, primarily due to the impact of small particulate matter on public health. This research has concentrated on the development of a Deep Learning model designed to automatically process images obtained from UAVs. The goal is to enhance the efficiency and precision of fire control efforts. In the past, satellite images were utilized for data collection, but their limited resolution and inability to provide real-time data posed challenges. The study's findings reveal that the use of a pre-trained model, specifically Densenet121, has the potential to deliver highly accurate results and effectively pinpoint the location of fires in aerial imagery.
Keywords - UAV, Deep Learning, Classification, Forest Fire, Automate Detection