基于面部信息的挖掘机驾驶员疲劳特征提取

Excavator driver fatigue feature extraction based on facial information

针对挖掘机驾驶员疲劳驾驶导致操作事故问题,文章提出一种基于面部信息的挖掘机驾驶员疲劳特征提取方法。首先采用基于AdaBoost算法对人脸和眼部区域进行定位,根据“三庭五眼”的脸部特征定位嘴部;其次用大津法二值化眼部图像,并根据其垂直积分投影判断眼睛状态;最后采用基于Canny的嘴部边缘检测方法获取嘴部似圆度判断嘴部状态。实验表明,该文提出的方法能准确获取驾驶员闭眼和打哈欠的次数,疲劳特征为后续疲劳检测提供了依据。

Aiming at the problem of operation accident caused by fatigue driving of excavator driver,an excavator driver fatigue feature extraction method based on facial information is presented.Firstly,the face and eye regions are located based on AdaBoost algorithm,and the mouth is positioned according to the facial features of“three courts and five eyes”.Then,the eye image is binarized by Otsu algorithm,and the eye state is judged by its vertical integral projection.Finally,the mouth edge detection method based on Canny algorithm is used to obtain the mouth roundness to judge the mouth state.Experiments show that this method can accurately obtain the times of eyes closure and yawning of the driver,and the fatigue characteristics provide a basis for subsequent fatigue detection.