驾驶员疲劳检测中眼睛特征提取算法研究
|
摘要
针对当前驾驶员疲劳检测过程中眼睛特征提取方法鲁棒性较差,且定位不准的问题,提出了一种驾驶员疲劳检测中眼睛特征提取方法。首先通过基于图像增强的ASM算法对人脸特征进行定位,利用模型中的定位点提取眼睛区域,在眼睛初步定位的基础上,分别利用加权混合投影算法和多特征融合的方法对内外眼角进行精确定位;然后根据内外眼角确定瞳孔区域,采用基于灰度分布的加权积分投影算法对瞳孔中心进行粗定位,利用形态学原理对瞳孔中心进行校正,从而得到瞳孔的精确位置;最后根据准确定位提取眼睛中点的上下眼睑点,从而获取眼睛开度的大小。试验表明,该算法能准确定位瞳孔和眼角位置,并利用定位点正确提取眼睛特征。
In this paper,a method of eye feature extraction in driver fatigue detection is proposed to solve the problem that the feature extraction method is poor and the location is not accurate.Firstly,ASM algorithm based on image enhancement of facial feature localization,eye region extraction using location model,based on the preliminary set eyes,respectively by using the weighted hybrid projection algorithm and multi feature fusion method to locate the corners; then according to the determined pupil area inside and outside corners,using the weighted integral projection the algorithm of gray distribution based on pupil center of the coarse positioning of the pupil center were corrected by using morphological principle,so as to obtain the precise position of the pupil; finally,according to the accurate positioning of two eye corner extraction point on the lower eyelid eyes point to get eye opening size.In this paper,experiments are carried out to test the above methods.The results show that the algorithm can accurately locate the pupil and eye corner,and use the location point to extract the eye features.
In this paper,a method of eye feature extraction in driver fatigue detection is proposed to solve the problem that the feature extraction method is poor and the location is not accurate.Firstly,ASM algorithm based on image enhancement of facial feature localization,eye region extraction using location model,based on the preliminary set eyes,respectively by using the weighted hybrid projection algorithm and multi feature fusion method to locate the corners; then according to the determined pupil area inside and outside corners,using the weighted integral projection the algorithm of gray distribution based on pupil center of the coarse positioning of the pupil center were corrected by using morphological principle,so as to obtain the precise position of the pupil; finally,according to the accurate positioning of two eye corner extraction point on the lower eyelid eyes point to get eye opening size.In this paper,experiments are carried out to test the above methods.The results show that the algorithm can accurately locate the pupil and eye corner,and use the location point to extract the eye features.
引文
[1]HU S Y,ZHENG G T.Driver drowsiness detection with eyelid related parameters by support vector machine[J].Expert System with Applications,2009,36:7651-7658.
[2]张伟.基于机器视觉的驾驶人疲劳状态识别关键问题研究[D].清华大学,2011.
[3]ALAN L Y,PETER W.Feature extraction from face using deformable templates[J].International Journal of Computer Vision,1992,99-111.
[4]谭华春,章毓晋,李睿.基于角点特征的眼睛轮廓提取[J].中国图像图形学报,2007,12(7):1224-1229.
[5]黄万军,尹宝才.三维可变形模板的眼睛特征提取[J].计算机研究与发展,2002,4:495-501.
[6]ZHONG L Z,JIE Y.A robust method for eye features extraction on color image[J].Pattern Recognition Letters,2005,26,2252-2261.
[7]孙艳蕊,田书贞.基于眼角精确定位的实现估计[J].东北大学学报,2014,35(6):780-784.
[8]潘林,魏丽芳.改进遮挡条件下瞳孔检测方法[J].中国图像图形学报,2012,17(2):229-235.
[9]王先梅,杨萍,王志良.多姿态眼球中的瞳孔定位算法[J].计算机辅助设计与图形学报,2011,23(8):1427-1432.
[10]夏海英,严国萍.加权方差投影在眼角定位中的应用[J].中国图像图形学报,2011,16(2):273-277.
[11]CUI X,YING Z,ZENG W.Semantic feature extraction for accurate eye corner detection[C].IEEE International Conference on Pattern Recognition(ICPR),2008.
[12]JOSE J,JUAN J,CERROLAZA B.Evaluation of accurate eye corner detection methods for gaze estimation[J].Journal of Eye Movement Research,2013,7(3):1-8.
[13]LIN Y,LIN R,LIN Y,et al.Real-time eye-gaze estimation using a low-resolution webcam[J].Multimedia Tools and Applications,2013,65(3):543-568.
[14]ZHANG W,WANG Z,XU J,et al.A method of gaze direction estimation considering head posture[J].International Journal of Signal Processing,Image Processing and Pattern Recognition,2013,6(2):103-112.
[15]刘春生,常发亮,陈振学,等.基于Gabor滤波和K-mediods聚类分析的人眼检测和瞳孔定位[J].光电子,2013,24(8):1606-1612.
[2]张伟.基于机器视觉的驾驶人疲劳状态识别关键问题研究[D].清华大学,2011.
[3]ALAN L Y,PETER W.Feature extraction from face using deformable templates[J].International Journal of Computer Vision,1992,99-111.
[4]谭华春,章毓晋,李睿.基于角点特征的眼睛轮廓提取[J].中国图像图形学报,2007,12(7):1224-1229.
[5]黄万军,尹宝才.三维可变形模板的眼睛特征提取[J].计算机研究与发展,2002,4:495-501.
[6]ZHONG L Z,JIE Y.A robust method for eye features extraction on color image[J].Pattern Recognition Letters,2005,26,2252-2261.
[7]孙艳蕊,田书贞.基于眼角精确定位的实现估计[J].东北大学学报,2014,35(6):780-784.
[8]潘林,魏丽芳.改进遮挡条件下瞳孔检测方法[J].中国图像图形学报,2012,17(2):229-235.
[9]王先梅,杨萍,王志良.多姿态眼球中的瞳孔定位算法[J].计算机辅助设计与图形学报,2011,23(8):1427-1432.
[10]夏海英,严国萍.加权方差投影在眼角定位中的应用[J].中国图像图形学报,2011,16(2):273-277.
[11]CUI X,YING Z,ZENG W.Semantic feature extraction for accurate eye corner detection[C].IEEE International Conference on Pattern Recognition(ICPR),2008.
[12]JOSE J,JUAN J,CERROLAZA B.Evaluation of accurate eye corner detection methods for gaze estimation[J].Journal of Eye Movement Research,2013,7(3):1-8.
[13]LIN Y,LIN R,LIN Y,et al.Real-time eye-gaze estimation using a low-resolution webcam[J].Multimedia Tools and Applications,2013,65(3):543-568.
[14]ZHANG W,WANG Z,XU J,et al.A method of gaze direction estimation considering head posture[J].International Journal of Signal Processing,Image Processing and Pattern Recognition,2013,6(2):103-112.
[15]刘春生,常发亮,陈振学,等.基于Gabor滤波和K-mediods聚类分析的人眼检测和瞳孔定位[J].光电子,2013,24(8):1606-1612.