【摘要】 疲劳驾驶是导致交通事故的主要原因之一，为了降低交通事故对人们生命财产的危害，本文采用分层梯度方向直方图(pyramid histogram of oriented gradients,PHOG)算法进行人脸识别和关键点检测，提出一种多特征疲劳特征因素的疲劳驾驶检测方法，结合OpenCV对人面部的眼、嘴以及头部空间姿态坐标点进行定位，设定眨眼、哈欠及点头的疲劳阈值，根据PERCLOS准则进行疲劳判定，最后采用朴素贝叶斯算法综合以上疲劳特征因素进行疲劳预测．实验数据表明，PHOG算法在各种复杂环境下的准确率均达到95%以上，具有很好的稳定性和抗干扰能力．更多还原

【Abstract】 Fatigue driving is one of the main causes of traffic accidents. In order to reduce the harm traffic accidents to people’s lives and properties,in this article we use pyramid histogram of oriented gradients(PHOG)algorithm for face recognition and key point detection,and propose a fatigue driving detection method based on multi-characteristic factors. In our proposed method,combined with OpenCV,the coordinate points of human eyes,mouth and head posture are first located,and then the eye,mouth and head posture are detected according to PERCLOS criterion to determine fatigue. Finally,the Naive Bayes algorithm is used to integrate the above fatigue characteristic factors for fatigue prediction. Our experimental data show that the accuracy of PHOG algorithm in various complex environments is more than 95%,which suggests that PHOG algorithm has good stability and anti-interference ability.更多还原