面部识别在检测疲劳驾驶的应用研究

Research on the Application of Face Recognition in Detecting Fatigue Driving

面部疲劳信息收集与报警能够有效降低驾驶员疲劳驾驶导致交通事故的概率。疲劳状况下,驾驶员的决策认知机能下降,比起平时驾驶习惯会出现在同一方向注视时间过长、眼动频率降低、决策执行时间延长的情况,按照PERCLOS准则,眨眼频率、打哈欠程度在判断驾驶员疲劳驾驶上发挥重要作用。本文提出在驾驶员非疲劳状态下自学习基于驾驶员个人习惯的行车状态面部状态,对模型进行训练得出阈值数据,更能科学地根据驾驶人的习惯和个人特殊的特征,人性化地优化面部识别在减少疲劳驾驶应用中的检测模型。

The collection and alarm of facial fatigue information can effectively reduce the probability of traffic accidents caused by drivers'fatigue driving.Under the condition of fatigue,drivers'decision-making cognitive function decreases.Compared with their usual driving habits,they will spend too much time looking at the same direction,reduce their eye movement frequency,and shorten their decision-making execution time.According to PERCLOS guidelines,blink frequency and degree of yawning plays an important role in judging drivers'fatigue driving.This paper proposes to self-learning the driving state facial state based on the driver's personal habits under the non-fatigue state of the driver,train the model to obtain threshold data,more scientifically according to the driver's habits and personal special characteristics,and humanize the facial recognition to optimize the detection model in the application of reducing fatigue driving.