基于AdaBoost算法的疲劳驾驶检测系统研究

为提高行车安全性,结合驾驶员人眼开度特征开展了疲劳驾驶检测系统研究,提出采用AdaBoost算法实现人脸及人眼的定位方法。驾驶员视频图像在动态直方拉伸和训练好的AdaBoost算法分类器作用下实现人脸、人眼定位;对获得的眼部图像进行边缘检测及轮廓提取,计算轮廓内像素点得到眼睛的开度信息;利用PERCLOS算法计算一段时间内驾驶员的闭眼帧数及眨眼频率比例,获得驾驶员疲劳状态。试验结果表明,该系统具有较好的抗环境干扰和实时性,能准确地完成疲劳判断。

Research on Fatigue Driving Detection System Based on AdaBoost Algorithm

In order to improve the driving safety,this paper studies the fatigue driving detection system based on the driver's eyes opening characteristics,and proposes the AdaBoost algorithm to realize the positioning method of the human face and the human eye.Firstly,the driver video image is used to realize the human face and eye positioning under the action of the dynamic histogram stretching and the trained AdaBoost algorithm.Secondly,the edge detection and contour extraction are performed on the obtained eye image,and the pixel points in the contour are calculated to obtain the opening information of the eye.Finally,the PERCLOS algorithm is used to calculate the number of closed eye frames and the proportion of the blink frequency of the driver over a period of time to obtain the driver's fatigue state.The test results show that the system has good anti-environmental interference and real-time performance,and can accurately complete the fatigue detection and judgment.