基于SVM的驾驶员疲劳检测研究

摘要

Driver fatigue is an important factor that causes traffic accidents. It can prevent traffic accidents to install a fatigue detection system in the car. In the real conditions, the driver's head and eyes are in motion, this makes the fatigue feature extraction become more difficult. External interference and the impact of lighting also make it a challenging problem to determine accurately the driver's fatigue. A method to detect driver fatigue with support vector machine is introduced. First, the video of driver's head is captured, and then the video images are processed, features of the eyes, mouth, nodding frequency of head are extracted, at last, support vector machines are used to determine the driver's fatigue status. The experiment results show that fatigue detection accuracy rate reached 97.80％, and the method is suitable for driver fatigue detection.%疲劳驾驶是导致交通意外的一个重要原因,在车上装一个疲劳检测系统有助于预防交通事故的发生.现实条件下,司机的头和眼睛是不断运动的,使得疲劳特征提取变得比较困难.再加上外部干扰和光线条件的影响,准确判断司机的疲劳状态是一个具有挑战性的问题.介绍了一种利用支持向量机检测驾驶员疲劳状态的方法.首先采集驾驶员的头部视频,然后对视频图像进行处理,提取眼睛、嘴的视觉特征和点头频率变化情况,最后利用支持向量机依据这些特征来判断司机的疲劳状态.通过模拟实验,疲劳检测的准确率达到97.80%,表明该方法适合于驾驶员的疲劳检测.