基于多特征的疲劳驾驶安全检测  

Fatigue Driving Safety Detection Based on Multi-Feature

作　　者：刘晨熙 刘芬[1] 

机构地区：[1]天津职业技术师范大学电子工程学院,天津

出　　处：《计算机科学与应用》2023年第4期680-688,共9页Computer Science and Application

摘　　要：针对目前因疲劳驾驶导致的车祸比例占总车祸比例的极大一部分的情况,故本文设计出一种基于改进的Yolov5算法的疲劳驾驶检测系统。Yolov5算法的疲劳检测是指通过对车辆内部安装摄像头等图像传感器来获取驾驶员的人脸和面部特征,并通过机器视觉中的面部特征提取等方式获取驾驶员眼部以及嘴部状态,从而实现对驾驶员疲劳状态的分析判断。基于Yolov5算法的检测方法成本低,不需接触、检测方便,能够准确地分析出驾驶人的疲劳状况,可以极大地降低因疲劳驾驶而导致的事故发生概率。Aiming at the fact that the proportion of car accidents caused by fatigue driving accounts for a large part of the total number of car accidents, a fatigue driving detection system based on the improved Yolov5 algorithm is designed in this paper. The fatigue detection of the Yolov5 algorithm refers to the acquisition of the driver’s face and facial features by installing image sensors such as cameras in the vehicle, and the driver’s eye and mouth status through facial feature extraction in machine vision, so as to realize the analysis and judgment of the driver’s fatigue state. Based on Yolov5 algorithm, the detection method has low cost, no contact, convenient detection, can accurately analyze the fatigue of the driver, and can greatly reduce the probability of accidents caused by fatigue driving.

关 键 词：Yolov5 深度学习 疲劳检测 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]