基于PERCLOS的疲劳驾驶检测系统的实现

摘要

本文阐述了PERCLOS测评驾驶疲劳的机理。介绍了疲劳识别的算法，该系统首先利用肤色模型粗略检测出入脸区域，并根据驾驶员面部几何特征，准确定位出眼睛；然后根据眼睛区域白色像素的多少及闭眼持续时间对驾驶员状态做出判断；最后应用MicrosoftVisualC＋＋6．0在WindowsXP操作平台上，对疲劳检测判定进行实现。实验表明：该系统可以准确定位眼睛并判定驾驶员疲劳状态，其准确率为94．44％。%This paper expounds the mechanism on PERCLOS detecting and evaluating fatigue driving. It introduces fatigue recogni- tion algorithm. This system firstly detects the face region roughly using skin - color model, and the drivers'eyes are located exactly by face geometry features. Secondly, a judgment of the drive's fatigue condition is made according to the number of white pixels in the eye area and the duration of driver's eyes closing. Finally, on the Windows XP operating platform applying Microsoft Visual C + + 6.0, it implements fatigue detection judgment. Experiments show that the system can accurately locate eyes make a judgment of the drive's fatigue condition, the precision rate is 94.44%.