Real time Driver Fatigue Detection Method for Advanced Driver Assistance System

摘要

Fatigue driving is one main cause for most serious traffic accidents,eye's status is the key clue to judge whether the driver is drowsy or not. This paper proposes a practical real time algorithm which is applied in the on-board system to detect eye's status with an infrared camera. Face detection method is first applied to localize the eyes of the driver,after extracting the eye region,eyelids status can be detected. An alarm rule is made in term of the eyelid status. The algorithm is tested in real scene in the day and at night,and with an on-board embedded Driver Fatigue Detection System(DFDS),which is a subsystem of Advanced Driver Assistance System(ADAS), the algorithm can provide above 92% accuracy and less than 1.5s react time in real driving scenes,it is available for real time use.