基于Adaboost和LBP算法的实时疲劳驾驶预警算法

Real-time Fatigue Driving Warning Algorithm Based on Adaboost and LBP Algorithm

为了有效减少因疲劳驾驶而导致的交通事故,提出一种基于Adaboost和局部二进制特征(local binary patterns,LBP)算法的疲劳驾驶预警算法。算法采用Canny边缘检测算法对图像帧进行轮廓检测,并消除原图像噪点,根据亮度条件分别处理图像帧。在日间环境状态下,根据Adaboost算法识别出眼部和嘴部位置及开合状态。夜间环境状态下,采用相邻帧间差分算法确定眼部的矩形特征,采用LBP算法来处理嘴部的开合状态。同时,为了评估眼部状态信息和嘴部状态信息对驾驶员是否处于疲劳驾驶状态的比重,文章还进行了实验比对。实验结果表明:当嘴部状态信息权重β为0.44时,系统判别精度最高,最高精度可达93.4%。

In order to effectively reduce traffic accidents caused by fatigue driving,a fatigue driving early warning algorithm based on Adaboost and local binary patterns(LBP)algorithms is proposed.The Canny edge detection algorithm is used to detect the contour of the image frame and eliminate the original image noise.The image frames are processed separately according to the brightness conditions.In the daytime environment,the Adaboost algorithm is used to identify the position of the eyes and mouth for the opening or closing state.In the night environment,the method of difference between adjacent frames is used to determine the rectangular feature of the eye,and the LBP algorithm is used to process the opening and closing state of the eye.At the same time,in order to evaluate the proportion of eye state information and mouth state information to determine whether or not the driver is in a fatigued driving state,this paper conducts an experimental comparison.The experimental results show that when the weight of the mouth state informationβis 0.44,the system has the highest discrimination accuracy,with the highest accuracy reaching 93.4%.