Automatic driver fatigue detection based on visual computing

Abstract

In the context of the growing e-commerce sector, the complexity of supply chains has surged, placing heightened demands on drivers facing increased fatigue. This is particularly critical for timely deliveries of perishable goods and time-sensitive services. To address this challenge and enhance transportation efficiency while mitigating road accidents, a driver fatigue detection system is essential.

This dissertation explores the integration of real-time monitoring of driving behavior, utilizing RGB cameras to detect signs of fatigue such as eye closure, yawning, and head position. The system issues warnings through an in-car display to prompt timely driver response. Notably, test results demonstrate a robust 97.8% accuracy in detecting eye closure. Future work could refine alert mechanisms to correct driver behavior more efficiently and add add infrared cameras to the system for easy detection in the dark, further optimizing the proposed fatigue detection system.