Nighttime High Beam Detection for Autonomous Driving Using YOLOv12

The rapid advancement of autonomous driving technology imposes stringent requirements on nighttime driving safety. High beam detection in nighttime scenarios has become a critical and challenging task for autonomous driving systems. However, existing methods primarily focus on third-person surveillance perspectives, which fail to meet the practical detection needs from a first-person viewpoint required by autonomous vehicles. To address this issue, this paper constructs a first-person nighttime high beam dataset and conducts experiments and result analysis using the YOLOv12 model. The proposed approach provides both data and methodological support for enhancing nighttime perception capabilities in autonomous driving systems.