基于眼睛和嘴巴状态的驾驶员疲劳检测算法

DRIVER FATIGUE DETECTION ALGORITHM BASED ON THE STATES OF EYES AND MOUTH

针对列车驾驶员疲劳检测问题,提出一种基于人眼和嘴巴状态的驾驶员疲劳检测算法。首先采用改进的AdaBoost算法精确定位驾驶员脸部区域。然后通过模板匹配定位人眼,并根据人脸的几何特征定位嘴巴。最后计算每一帧图像的PERCLOS(per-cent eyelid closure)参数和嘴部动作频率,统计单位时间内双参数与对应阈值的关系,作为判断驾驶疲劳的依据。实验结果表明,在正常光照下,综合眼睛和嘴部信息,比采用单参数检测算法减少了误判、漏判的概率,具有较高的准确性和鲁棒性。

Aiming at the problems of train drivers fatigue detection,we propose a fatigue detection algorithm which is based on the states of eyes and mouth of drivers.First,the improved AdaBoost algorithm is used to accurately locate the face area of drivers,then the human eyes are located through template matching,and the mouth is positioned according to the geometrical characteristics of human face.Finally the PERCLOS parameters of each frame of the image and the frequency of mouth movements are calculated,the relationship of double parameters within the unit time and the corresponding threshold is counted as the base of driving fatigue judgement.Experimental results show that to integrate the information of eyes and mouth reduces the probability of misjudgement and judgement leaks compared with the detection algorithm using single parameter under the condition of normal light.The method has good precision and robustness.