基于机器视觉的疲劳驾驶监测算法研究

Research on Fatigue Driving Monitoring Algorithm Based on Machine Vision

疲劳驾驶严重威胁驾驶安全,目前,车辆制造商常采用对驾驶时间计时或对车辆行驶姿态分析的方式来进行防疲劳驾驶提醒,存在适应性差、准确率不足的问题。为了解决这个问题,提出了一种基于机器视觉的疲劳驾驶监测算法:通过边缘直方图相关性匹配算法消除图像背景并提取出驾驶员图像,再利用肤色聚类实现对驾驶员面部提取,基于面部灰度分布特征完成眼睛部位定位分割,再计算眼睛区域图像的灰度直方图和标准差,最后,设定阈值来判断眼睛闭合、睁开的状态。如连续多帧视频里眼睛都处于闭合状态,则属于疲劳驾驶。实验表明,算法准确度高、计算量较小,能有效地监测到驾驶员的疲劳驾驶状态。

Fatigue driving poses a serious threat to driving safety,and the way of timing the driving time or analyzing the driving attitude of vehicles was often used by vehicle manufacturers to remind them of fatigue driving.However,these methods have the problems of poor adaptability and insufficient accuracy.Therefore,in order to solve this problem,it proposed a fatigue driving monitoring algorithm based on machine vision.Firstly,the image background was eliminated by the edge histogram correlation matching algorithm to extract the driver’s image,and then the driver’s face was extracted by skin color clustering.Then,the eye location and segmentation were completed based on the gray distribution characteristics of the face,and then the gray histogram and standard deviation of the eye area image were calculated.Finally,the threshold is set to determine whether the eyes are closed or open.If the eyes are closed in consecutive multi frame videos,it is considered to be fatigue driving.Experiments show that the algorithm has high accuracy and less calculation,and can effectively monitor the driver’s fatigue driving state.