基于Blackfin533的驾驶疲劳实时检测预警系统

相关调查显示,疲劳驾驶是机动车交通事故发生的一个重要原因,以Blackfin533型DSP处理系统为硬件平台,设计了一套嵌入式实时的疲劳检测预警系统,以减少交通事故的发生率;主要原理是首先通过摄像头实时地采集驾驶舱内的监控视频图像,然后通过肤色的快速图像处理算法来定位人脸和人眼,最后提出一种根据虹膜外接矩形的形状特征近似计算PERCLOS值来评价驾驶员的疲劳程度,并给出报警信息;结果显示该系统在室内的实验中取得了良好的精度及速度。

Embedded Driver Fatigue Real-time Detecting and Warning System Based on Blackfin533

Surveys show that driver fatigue is an important cause of combination-unit traffic crashes.We try to develop an embedding real-time deriver fatigue detecting and warning system to decline the ratio of traffic crashes.Our design will be based on the Blackfin535 hardware console.The main principle of this system is that: first of all,we retrieve video images of the inside of the cabin by using a camera,then locate the driver's face and eyes by using the image processing algorithms based on skin color.Finally we propose a new method to compute PERCLOS value approximately by using the features of the rectangle border of iris,and output the warning signals.The results of indoors experiments show good detection rate and precision.