Automatic Fighter Identification System Utilizing YOLOv11 and Data Augmentation

Rapid and accurate Identification Friend or Foe (IFF) is crucial in modern warfare to prevent fratricide. This paper proposes an automated IFF system using the YOLOv11 object detection algorithm. To address the scarcity of training data for adversarial assets like North Korean (KPAF) aircraft, we developed a high-quality synthetic dataset combining 1/72 scale models with Blender-based 3D rendering. A Physically Based Rendering (PBR) workflow replicated realistic metallic optical properties to bridge the domain gap. Validated on the NVIDIA Jetson Orin Nano edge platform, the YOLOv11 model achieved superior performance with an mAP50 of 0.95 at 34 FPS. Significantly outperforming YOLOv8 and YOLOv10, this research provides a robust foundation for deploying low-cost, real-time optical IFF systems in practical military applications.