摘要
提出一种基于Adaboost的实时算法,并应用于车载嵌入式系统。用红外光源和红外摄像头获取驾驶员的视频图像,对其疲劳状态进行监控。首先通过人脸检测定位驾驶员的人脸,然后提取人眼区域并对人眼闭合状态进行判断,基于PERCLOS标准制定了相应的预警机制,对潜在的疲劳驾驶进行判断并预警。该算法从PC移植到嵌入式平台并根据实验进行了优化,先后制作了多个车载嵌入式装置进行实车测试,达到了大于92%的准确率和少于1.5 s的判断响应时间。实验装置稳定可靠,可实际应用于营运车辆。
The authors proposed a real time driver fatigue detection algorithm based on Adaboost. The algorithm is applied in an on-board embedded system to monitor driver＇s status with infrared camera and active IR lamps. Face detection method is used to localize the eyes of the driver and the eye region is extracted to monitor the movement of eyelids. An alarm rule is designed based on the PERCLOS standard to detect drowsy driving. The algorithm is transplanted from PC to embedded platform and optimized by experiments. Several versions of on-board embedded system are designed and manufactured to test the algorithm, it can provide above 92% accuracy and under 1.5 s reaction time in real driving scenarios. The experiments show that the system is stable and it is available for commercial use.
出处
《北京大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2012年第5期719-726,共8页
Acta Scientiarum Naturalium Universitatis Pekinensis
基金
深港创新圈专项计划(SG200810140026A)资助
