摘要
传统的疲劳检测装置,在图像获取效率和识别准确率方面,均不能很好地满足实际需求。采用人脸识别的开源数据库,通过人脸追踪获得人脸图像,对人的面部进行特征识别,结合驾驶时长设计的疲劳驾驶检测系统,可判断驾驶者的疲劳状态。实验结果表明,该系统与无人脸追踪的疲劳检测算法对比,在图像获取效率和识别准确率上分别提升30%和10%。
The traditional fatigue detection devices cannot meet the actual needs in terms of image acquisition efficiency and recognition accuracy.The fatigue driving detection system designed in this paper uses an open source database of face recognition to obtain face images through face tracking,recognize the features of the human face,and judge the fatigue state of the driver based on the driving time.The experimental results showed that compared with the fatigue detection algorithm without face tracking,the system improves image acquisition efficiency and recognition accuracy by 30%and 10%,respectively.
作者
刘硕
贺丽丽
LIU Shuo;HE Li-li(Department of Physics and Electronic Engineering, Yuncheng University, Yuncheng 044000, China)
出处
《运城学院学报》
2021年第6期51-54,共4页
Journal of Yuncheng University
基金
运城学院科研项目(CY-2019022)
运城学院大学生创新创业训练项目(DC-2020035)。
关键词
图像处理
疲劳检测
人脸追踪
image processing
fatigue monitoring
face tracking
