Detection of fire zone by unmanned aerial vehicle and prediction of fire intervention with artificial intelligence

Today, forest fires are one of the leading disasters that threaten living habitats. The increase in global warming, deterioration of air quality, decrease in animal and plant diversity are the main damages caused by forest fires. In order to combat the threats posed by forest fires, all states develop various policies and make serious financial expenditures. It is essential to develop technological products used in this struggle and to find a faster and more precise solution to forest fires. In this thesis, an artificial intelligence-based unmanned aerial vehicle system was developed based on these problems. The developed system has autonomous flight and analysis capabilities by patrolling the potential fire zone. With the camera on the system, instant images are transmitted to the ground unit and processed. The processed images are searched for signs of fire and smoke. In the event of a possible fire, the system autonomously flies in the fire zone and performs missions to intervene in the fire. At the time of fire, wind speed, UAV speed, UAV altitude information is processed with artificial intelligence models and a horizontal shot prediction is made. As a result of the prediction, the moment when the fire extinguishing ball should be released while the UAV is in motion is calculated. The model predicts how far the resulting horizontal shot will travel on the X-axis. The end point of the path to be taken from the X axis is estimated for the fire center. Thus, it is aimed to drop the fire extinguishing cannon into the flaming area with an accurate shot.