摘要
本文提出了一种运用近红外线主动光源的图像采集方法,能有效地克服光照变化带来的影响;提出了一种基于AdaBoost算法的人眼检测方法,能快速准确地检测出图像中驾驶员的眼睛区域,同时给出了一种人眼开合度评价方法来判断驾驶员的眼睛状态,从而达到检测驾驶员疲劳的目的。实验结果表明,采用本文的方法能有效地检测出驾驶员的疲劳状态,可作为一种驾驶员辅助系统安装在车辆内来实时监测驾驶员是否疲劳,并提醒打瞌睡的驾驶者。
This paper employs an active lighting system by using near infrared LED, which produces good face images regardless of visible lights in the environment. A fast approach of eye detection is also presented based on AdaBoost algorithm, which can extract drivers＇ eye regions accurately. Moreover, this paper proposes an evaluation means of eye openness of drivers. The experimental results show that the proposed methods can effectively detect the fatigue status of drivers. The proposed monitoring system can be installed in cars as a driver assistance system in order to warn drowsy drivers in real time,
