YOLO-SAT: An Edge-Efficient UAV-Based Smart Mobility Framework for Real-Time Desert Waste Detection with Data-Centric and Model-Centric Optimization

The global waste crisis is escalating, with solid waste generation expected to increase tremendously in the coming years. Traditional waste collection methods, particularly in remote or harsh environments like deserts, are labor-intensive, inefficient, and often hazardous. This motivates smart mobility and logistical strategies that use UAV monitoring to expand coverage and enable targeted clean-up. Recent advances in computer vision and deep learning have opened the door to automated waste detection systems, yet most research focuses on urban environments and recyclable materials, overlooking organic and hazardous waste and underex-plored terrains such as deserts. In this work, we propose YOLO-SAT, an enhanced real-time object detection framework based on a pruned, lightweight version of YOLOv12 integrated with Self-Adversarial Training (SAT) and specialized data augmentation strategies. Using the DroneTrashNet dataset, we demonstrate improvements in precision, recall, and mean average precision (mAP), while achieving low latency and compact model size suitable for deployment on resource-constrained aerial drones. Benchmarking YOLO-SAT against state-of-the-art lightweight YOLO variants further highlights its balance of accuracy and efficiency. Our results validate the effectiveness of combining data-centric and model-centric enhancements for robust, real-time waste detection in desert environments, supporting UAV-assisted monitoring for targeted clean-up planning and operational resource allocation.