基于红外差频嵌入式疲劳驾驶监测系统的研究  

Study on embedded real-time driving fatigue monitoring based on infrared ADF

作　　者：刘志强[1] 张海水[2] 詹永照[2] 

机构地区：[1]江苏大学汽车与交通工程学院,江苏镇江212013 [2]江苏大学计算机科学与通信工程学院,江苏镇江212013

出　　处：《计算机工程与设计》2009年第9期2096-2097,2300,共3页Computer Engineering and Design

基　　金：国家科技支撑计划基金项目(2007BAK35B02);江苏省科技厅2005年高新基金项目(BG2005028)

摘　　要：针对疲劳驾驶监测的要求,通过对传统图像处理以及基于DSP图像处理方法的研究,提出了基于DSP红外差频法检测眼睛来进行疲劳监测的方法。通过DSP控制近轴和远轴红外灯的开关,利用红外光源下近轴亮瞳孔和远轴暗瞳孔进行图像差分来获取眼睛区域,平滑去噪二值化后进行特征提取,最后计算PERCLOS值来判断疲劳程度并决定是否报警。实验结果表明,该方法准确率较高、速度较快,可以满足非接触式、不同光照条件、实时监测的要求。Aiming at the request of the driving fatigue monitoring system,a driver fatigue detection method based on DSP is proposed,which utilizes ADF under IR light to detect eyes by using DSP based and common image processing method.The switch of inner and outer IR light is controlled by DSP,under the IR light the bright pupils and the dark pupils will be generated inner and outer the axis,repectively.It could gain the area of the eye from the frame generated by subtracting the dark pupils frame from the bright one.The characteristic parameter of eyes is captured with smoothing and high-low.Finally the PERCLOS value which decides the degree of tiredness is calculated out.The PERCLOS value also decides whether alarm is necessary or not.Experimental results show the fatigue detection rate is high and the speed is fast in this system.It can satisfy the demands of non-contact,different lighting conditions and real-time monitoring.

关 键 词：疲劳检测 数字信号处理 PERCLOS 差分图像 瞳孔 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]