High-Precision Multi-Class Object Detection Using Fine-Tuned YOLOv11 Architecture: A Case Study on Airborne Vehicles
Nasser S. Albalawi
Abstract
The widespread adoption of airborne vehicles, including drones and UAVs, has brought significant advancements to fields such as surveillance, logistics, and disaster response. Despite these benefits, their increasing use poses substantial challenges for real-time detection and classification, particularly in multi-class scenarios where precision and scalability are essential. This paper proposes a high-performance detection framework based on YOLOv11, specifically tailored for identifying airborne vehicles. YOLOv11 integrates innovative features, such as anchor-free detection and enhanced attention mechanisms, to deliver superior accuracy and speed. The proposed framework is tested on a comprehensive airborne vehicle dataset featuring diverse conditions, including variations in altitude, occlusion, and environmental factors. Experimental results demonstrate that the fine-tuned YOLOv11 model exceeds the performance of existing models. Additionally, its ability to operate in real-time makes it ideal for critical applications like air traffic management and security monitoring.