Automated Helmet Detection System Using RT-DETR for Real-Time Monitoring of Motorcyclist Safety

Abstract—The rising number of road accidents, especially those involving motorcyclists, underscores the urgent need for effective safety compliance solutions. This research proposes a real-time helmet detection alarm system powered by Detection Transformers (DETR) to enhance detection accuracy and opera- tional efficiency. The system is designed to identify motorcyclists who are not wearing helmets in real time and generate alerts to promote safety. DETR, an advanced deep learning model, utilizes the self-attention mechanism to capture complex and contextual relationships in image sequences, enabling precise helmet detection even under challenging conditions such as poor lighting or varied environmental factors [7]. The archi- tecture of the proposed system combines video feed analysis with DETR’s end-to-end object detection capabilities, ensuring minimal processing delay. Testing results reveal the system’s high performance, with impressive precision and recall rates in various scenarios, including urban roads and highways [8]. This solution can be customized to send alerts to relevant authorities or directly notify the rider via integrated signaling systems, thus potentially reducing violations and enhancing motorcyclist safety on the roads. Index Terms—Helmet detection, RT-DETR, Road safety, Ob- ject detection, Motorcycle helmet compliance, Video feed analysis