GAN-UAV-YOLOv10s: improved YOLOv10s network for detecting small UAV targets in mountainous conditions based on infrared image data

With the strong development of the field of computer science, deep learning methods have made strides in unmanned aerial vehicle (UAV) detection. However, targeting small infrared UAVs in mountainous conditions is still a challenge for the scientific community, causing many impacts from mountainous conditions such as thermal interference, interference due to complex terrain, and interference from weather. In this paper, we have proposed the UAV-YOLOv10s model based on the YOLOv10s model to detect small infrared UAV targets in mountainous conditions. Additionally, to improve the ability to detect small UAV targets, we proposed using a GAN network to augment data for small UAV targets. The experimental results show that the proposed model gives better results than the four newest models today: YOLOv8s, YOLOv9s, YOLOv10s and YOLOv11s. © The Author(s), under exclusive licence to Springer-Verlag London Ltd., part of Springer Nature 2025.