Detection of Safety Helmets Using YOLOv5s Model

With the increasing number of vehicles, these days the road accident is becoming very common in society. The manual detection of helmet is very time-consuming and requires a lot of manpower. In order to overcome the abovementioned issue, an automated technique is required detect whether the helmet is worn by the rider or not. This research proposes a helmet detection model for checking whether the riders are wearing helmets or not. The research utilizes the YOLOv5s deep learning object detection algorithm to identify rider's helmets. In this research work, various combinations of learning rates, levels, and epochs are explored to assess the model performance and identify the optimal model. Performance metrics such as F1 score, precision, recall, and Mean Average Precision (mAP) are computed to evaluate the effectiveness of the proposed system. To validate the experimental results, the proposed model is compared with the recent existing research works.