基于RBF网络的眼睛状态识别  

Recognition of Drivers' Eye State Based on RBF Network

作　　者：路玉峰[1] 王增才[1] 杨秀建[1] 

机构地区：[1]山东大学机械工程学院,济南250061

出　　处：《机电一体化》2007年第6期28-29,47,共3页Mechatronics

基　　金：教育部博士点基金(编号20060422011)

摘　　要：在基于视频分析的驾驶员状态检测过程中,对驾驶员眼睛状态的检测是关键一环。经研究发现,人在清醒、疲劳、睡眠时,眼的张开程度不同,并且这3种状态时眼睛区域二值化图像的水平投影有明显区别。提取眼睛区域二值化图像水平投影特征,利用RBF网络分类器识别驾驶员眼睛的状态。实验结果表明,方法可行,并取得满意效果。The detection of drivers＇ eyes state is a key part in measurement of drivers＇ state based on vision analysis. We found that the opening degree of drivers＇ eyes is different when they are sober, fatigue and drowsing. And the horizontal projections of three states eyes area binary images are distinctly different. The features of those projections were extracted from the images, and we may recognize the eyes states using RBF network. The experiments result that the method is effect and the results are encouraging.

关 键 词：眼睛状态 二值图像投影 特征提取 状态识别 RBF网络 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]