Drone aerial target detection based on YOLOv8

Abstract

Target detection in drone aerial photography is an important application scenario in the field of target detection. However, based on the characteristics of drones such as fast movement speed and high flight altitude, the detection target size of drones is small, the image scene changes greatly, and the real-time requirements are higher. This article uses the YOLOv8 model as a target detection method for drone aerial photography, and conducts comparative experiments on other classic target detection algorithms. The experimental results show that YOLOv8 has a better improvement effect compared to other target detection algorithms.