摘要
人眼定位是非接触式机车车辆乘务员实时疲劳检测中最基本的关键环节。针对机车车载视频序列进行疲劳检测时存在实时性、鲁棒性要求高的问题,结合计算机视觉和图像处理技术,提出一种多算法融合的人眼疲劳快速定位方法。利用基于Haar-like特征的Ada Boost人脸检测方法检测视频序列中乘务员人脸并设置目标跟踪区域,利用Camshift目标跟踪算法快速定位人脸;基于人脸和人眼的对称性等先验知识,提出比例缩减区域(PRA)的方法快速定位人眼。研究结果表明:该方法可以有效减少误检,适应复杂光照,并具有实时性和鲁棒性,为机车车辆乘务员疲劳实时检测提供理论和实践参考。
Eye location is the key link in the real-time fatigue detection of non contact locomotive drivers.In order to solve the problem of real-time performance and high robustness in locomotive-mounted video sequences,a new method of fast eye fatigue location based on multiple algorithm fusion was proposed by combining computer vision and image processing techniques.Firstly,it detected face in video sequence to use the method of AdaBoost based on Haar-like feature,setting target tracking area and using Camshift algorithm to track it.Secondly,the method of the Proportion Reduction Area was proposed to locate eye based on the prior knowledge of the symmetry of face and eye.Experimental results show that the method reduces the complexity and error detection,which can be adapted to the complex illumination,and has the advantages of real-time and robustness.It can provide theoretical and practical references for real-time fatigue detection of locomotive drivers.
作者
贺德强
刘卫
卢凯
肖琼
江洲
HE Deqiang;LIU Wei;LU Kai;XIAO Qiong;JIANG Zhou(College of Mechanical Engineering,Guangxi University,Nanning 530004,China;Nanning CRRC Rail Transit Equipment Co.Ltd,Nanning 530200,China)
出处
《铁道科学与工程学报》
CAS
CSCD
北大核心
2018年第9期2359-2366,共8页
Journal of Railway Science and Engineering
基金
国家自然科学基金资助项目(51165001)
广西科技攻关资助项目(1598009-6)
南宁市科技攻关资助项目(20151021)
