A YOLOR Based Visual Detection of Amateur Drones

The popularity of Unmanned Aerial Vehicles (UAV) has increased considerably among today’s technologies. Especially in the field of defense technologies, it has become very important to detect UAVs in military sheltered areas, border violations or any special area that needs protection. In this study, the recently released YOLOR was used to detect UAVs. The data set created from the UAV images was trained with the YOLOR model, and as a result of the experiments, it was experimentally proven that the YOLOR model has a 95.9% success rate and is better than other object detection models.