基于PERCLOS的驾驶员眼睛状态检测方法

利用驾驶员眼睛的状态判断驾驶员的疲劳状况是一种实时的、非接触式的方法。首先对眼睛进行精确定位,再根据眼睛面积的减小程度来判断眼睛的睁开或闭合状态,进一步通过记录眼睛闭合的次数、眼睛持续闭合的时间来检测驾驶员是否处于疲劳状态。本文采用肤色分割、去除与边界连通的区域、逐步减小眼睛区域候选范围的方法来定位人眼,再计算眼睛区域的面积,眼睛闭合时间,此方法不受复杂背景影响,对各种肤色、各种姿态的人脸均适用。

Driver's Eye State Detection Method Using PERCLOS

Using driver's eye state to judge driver's fatigue is a real-time and untouched method.After locating eyes exactly,judge driver's eye state is open or closed by counting eye area and area's reduction extent.Secondly,note eye's closure times and the continual closure time to detect driver's fatigue.By skin color segmentation,removing the object connected to the border and minish-ing the range of eye area gradually to locate driver'eye,and then compute the eye closure times and time.This method can't be influenced by complex background,and it is suitable to kinds of faces of different skin color and guise.