摘要
文章采用一种基于眼睛闭合度及打呵欠来检测驾驶员疲劳的方法,在YCrCb颜色空间中利用高斯模型进行肤色检测得到人脸的区域,在人脸灰度二值化图中利用五官几何结构的先验知识粗略定位人眼,利用区域生长和形态学运算得到人眼轮廓并计算眼睛的闭合度;检测嘴唇时利用唇色最佳阈值大致确定嘴唇位置,在此基础上通过人脸灰度值特征精确定位嘴唇,然后通过嘴张开程度判断驾驶员是否打呵欠;最后基于2个特征对驾驶疲劳进行判决,实验证明这种方法对驾驶疲劳检测具有较好的效果。
A method to detect driving fatigue based on the features of eyes and yawning is proposed. First, the face area is detected and located by using the Gaussian model in YCrCb color space. Then the facial gray image is linearized, and in the binary image, human eye regions are robustly located un- der the geometric constraints. By using the region growing and the morphological operating, the eyes positioning is accurately performed. Accordingly, the closure of eyes is calculated. Then the candidate lip area is located according to the best threshold of color space and the facial gray value feature. The degree of mouth opening shows whether the driver yawns. Finally, the driving fatigue is decided based on two facial features. The detection result of driving fatigue is improved because of the combi- nation of the features of eye and yawning frequency.
出处
《合肥工业大学学报（自然科学版）》
CAS
CSCD
北大核心
2013年第4期448-451,共4页
Journal of Hefei University of Technology：Natural Science
基金
合肥工业大学重大预研基金资助项目(2011HGZL0001)
关键词
疲劳驾驶
肤色
唇色
疲劳判决
灰度值特征
fatigue driving
skin color
lip color
fatigue decision
gray value feature
