基于DSP的疲劳驾驶预警系统的设计:本文描述了一个基于TI公司高速图像处理芯片TMS320DM642的疲劳检测预警系统,用以判定驾驶员的疲劳程度,降低交通事故。本系统采用非接触方法,先通过CCD摄像头拍下驾驶员的面部图像序列,然后检测出脸部,从而定位眼睛,再利用PERCLOS判定疲劳状态。本系统还结合免费的开源计算机视觉库OpenCV来加快开发速度和降低开发成本。Abstract: This paper describes a drowsiness detection system based on High-speedDSP—TMS320DM642 from TI Inc .It can determine the driver’s level of attention, and reducethe number of traffic accident. This system uses the untouched method. It gets video sequences ofa driver from the CCD camera, and than detects the face, eyelid. Finally, it uses the PERCLOScriterion to determine the state of drowsiness. In addition, it is programed by Open ComputerVersion (OpenCV), which can speeds up the development and reduces the cost.Keywords: face detection, eye localization, PERCLOS, OpenCV