<![CDATA[This study proposes a real-time motorcyclist drowsiness detection system using a camera-based approach to monitor eye blinks and assess rider awareness. The method integrates YOLO V5 for Region of Interest (ROI) extraction and Channel and Spatial Reliability Tracking (CSRT) for precise eye tracking on an inverted Cartesian plane. CSRT, leveraging CSR-DCF (Discriminative Correlation Filter), ensures robust movement detection. Blink frequency and interval analysis determine drowsiness levels, triggering timely alerts to enhance road safety. Designed for embedded deployment, the prototype utilizes a Raspberry Pi 4B, efficiently processing images via the Raspberry Pi Camera Module 2 NoIR. Experimental results confirm the system’s feasibility for real-time drowsiness detection on lightweight hardware, contributing to the advancement of intelligent rider safety technologies.]]>
Copyrights © 2025
