基于多疲劳信息融合的车载疲劳检测系统的研究和实现

Research and Implement of Driver Fatigue Detection Based on Multi-information Fusion

摘要:

疲劳驾驶是目前道路交通的一个重要安全隐患,对车载疲劳驾驶系统的研发具有重要的应用价值和广阔的市场前景；目前存在的疲劳监测系统普遍存在成本高,可靠性不足,使用不够便捷等缺陷；通过在安卓平台上采用计算机视觉的途径进行开发和实现,将大大降低系统的成本和使用复杂度；通过多种疲劳特征融合的方法对疲劳状态进行综合性判断,可以有效地增加系统的准确性和可靠性；采用优化的二叉树支持向量机多分类算法能够使得特征融合的过程具有准确性和速度上的优势；在对该疲劳检测系统的相应测试中也获取了良好的实验结果。

Abstract:

Currently, the driver fatigue is the great concern of road transportation safety. The research for fatigue driving detection is extremely practical and has a promising market future. Most of the existing fatigue detection system have the defect of high cost, lacking of reliability and inconvenient. The computer vision on Android platform is used for implementation, which will greatly reduce the cost and complexity. The fatigue detection mechanism based on multi-information fusion is adopted, which enhances the accuracy and robustness of the system. The optimized BT-SVM is used which can provide the advantage on the speed and accuracy. Good experimental results are obtained on related tests of the system.