This dataset was prepared for our 2019 project, “Amateur Drone Detection and Tracking.” It includes more than 4,000 pictures of amateur drones, such as DJI Phantom models. Additionally, the dataset contains non-drone, drone-like “negative” objects to enhance the robustness of detection models. Furthermore, it has been used with Yolov2-tiny and Yolov3-voc versions, making it generally suitable for working with the Yolo architecture and darknet framework. By including a variety of drone and non-drone images, this dataset provides a comprehensive resource for training and evaluating object detection models in aerial imagery, supporting advancements in drone detection technology.