부산대학교 도서관

Riding without a helmet is one of the leading causes of injuries and deaths in motorcycle accidents. In a country like Vietnam, the number of motorcycles on the roads is very high, making it difficult to monitor and preserve the safety of the riders. This research aims to propose a method for detecting motorcycle riders who are not wearing helmets using videos from surveillance cameras along the roads, which can help enhance law enforcement and manage the safety of riders. The proposed method utilizes the state-of-the-art object detection algorithm YOLOv5 to detect objects such as helmet, non-helmet, and rider. Next, it determines whether or not riders are wearing helmets in the post-processing step. Finally, the results showed that the detection model has an mAP (mean Average Precision) of around 98% and the proposed approach is able to identify the motorcycle riders who are not wearing helmets precisely.