Fatigue Driving Decttion Based On Facila Features And Head Posture

With the rapid development of economy and technology,in order to facilitate travel,private cars have become a must-have for everyone.Global statistics show that traffic accidents are the first cause of abnormal death.A National Highway Traffic Safety Administration(National Highway Traffic Safety Administration)research shows that a large proportion of traffic accidents in the United States are caused by driver fatigue driving.Therefore,in order to protect the life safety of drivers and passers-by,it is an urgent task to study the real-time detection system for fatigue driving,and it also has extraordinary significance.This paper conducts theoretical research on several schemes of fatigue driving.In order to improve driving safety,by comparing several fatigue detection algorithms that have been proposed at home and abroad,and combining their advantages,a self-based method based on head posture and facial features is proposed.Adaptive fatigue driving detection system,this system can detect the driver's fatigue in real time.The fatigue driving detection system consists of two subsystems,the first subsystem is the facial feature detection system,and the second subsystem is the head posture detection system.The content of this study in this article is divided into the following six points:(1)Fatigue detection methods for nearly 20 years are put together to put forward the fatigue detection system in this paper.(2)The Haar-like feature and the Adaboost algorithm are studied.The algorithm is used to determine whether the driver is present.Experiments show that this algorithm can quickly and accurately detect faces in the cockpit.At the same time,the EAR algorithm is studied.This algorithm uses the Dlib model to extract the 12 feature points of the eye and calculate the aspect ratio of the eye.If the aspect ratio of the eye suddenly drops and then rises quickly,it is judged as a blink.The blink detection rate can reach 99.2%.(3)A new BFR(Blinking Frequency Ratio)algorithm is proposed and applied to the field of fatigue driving for the first time.The BFR algorithm is a fatigue detection algorithm based on blinking time series,which studies normal blinking frequency and fatigue The average mapping relationship of blink frequency.The algorithm first collects the number of blinks of the driver in unit time,and generates blink frequency samples.During driving,the calculated BFR threshold is used to detect whether the driver has visual fatigue.If the number of blinks exceeds the BFR threshold,an alarm is given.(4)The yawn detection based on the MAR algorithm is studied.This algorithm also uses Dlib to extract 6 feature points of the mouth and calculate the aspect ratio of the mouth.If the aspect ratio rises above the MAR threshold and then drops rapidly,it is judged as A yawn action.The detection rate of yawn is 100%.(5)The head posture detection algorithm based on the Rodriguez rotation formula is studied,and the fatigue evaluation is performed using the P80 criterion,which is derived from the classic fatigue driving algorithm PERCLOS.Based on a pitch angle of 0 degrees and a roll angle of 0 degrees,if the deviation exceeds 20% of the maximum amount,it is determined that the driver's head is in a fatigue state.(6)Finally,the algorithm uses the Raspberry Pi platform for simulation,which can perform real-time detection.If the driver is detected to be in a fatigue state,the buzzer can be controlled to make a sound to realize real-time warning of fatigue driving.