Researchers at Kherson National Technical University have developed an innovative approach to detecting and monitoring forest fires using drone technology.
The system utilizes optical and infrared cameras, along with LiDAR, to accurately identify the location and spread of fires.
The system is designed to use various types of sensors. Optical cameras identify smoke, while infrared cameras measure temperature.
The system employs a specialized four-channel data processing scheme: smoke, heat, vegetation, and fire.
The system uses a YOLOv8N neural network model to enhance the efficiency of data processing and the accuracy of fire detection.
The system was tested on a computer with a Pentium i7-10700 processor, and it was found that the accuracy of forest fire front recognition improved by 21% compared to the use of multiple drones.
The system also measures important parameters of the fire: the intensity of heat radiation, the speed of fire spread, and the color of the flame.
The research results were published in the scientific journal "Science and Technology Today."