Performance Analysis of YOLOv11 and YOLOv8 for Mixed Traffic Object Detection under Adverse Weather Conditions in Developing Countries
A new study introduces YOLOv11 Nano, an updated iteration of the YOLO object detection series, and benchmarks it against YOLOv8 Nano. The research evaluated their performance on a fused dataset combining Indian Driving Dataset and Berkeley Deep Drive Dataset, focusing on mixed traffic scenarios under adverse weather conditions like rain and low light. YOLOv11n demonstrated a 3.2% improvement in precision, achieving a mAP@50 of 46.6%, while also reducing computational load by 22% and maintaining real-time inference speeds of 70.9 FPS on a Tesla T4 GPU. AI
IMPACT YOLOv11 Nano offers improved accuracy and efficiency for object detection, potentially enhancing autonomous driving systems in challenging conditions.