Design Of Fatigue Detection System Based On Eye And Head State

In recent years,the incidence of traffic accidents increases day by day.Fatigue driving is one of the important causes of traffic accidents,which brings great economic loss and personal injury to people.The serious consequences of fatigue driving are often ignored on the premise of the prohibition of drunk driving.Therefore,how to effectively prevent and reduce the incidence of fatigue driving is of great practical significance.Based on the analysis of domestic and international research on driver fatigue,through the changing states of eyes and head(left deviation,right deviation,looked up and nod),achieving a fatigue testing system.In order to further improve the accuracy of fatigue testing.In the course of driving,the system realizes the real-time monitoring of the driver and gives an alarm to the driver timely and effectively.The main contents are as follows:(1)Considering the influence of illumination and noise during the driving process,we use median filter and histogram equalization to compensate illumination,and ensure the accuracy of subsequent face detection by simple pretreatment.This thesis analyzes the commonly used face detection method of Adaboost and the Seetaface detection method used in this paper.The two methods are compared and analyzed in terms of speed,accuracy and head deflection.Finally,Seetaface is selected as a better detection algorithm.(2)On the basis of face detection,face tracking and optimization are carried out to improve the detection speed.The feature points are located roughly according to the geometric features,and the face alignment module is used to locate the facial feature points accurately,including eyes,nose and two corners of mouth.For the extracted human eye features,adopts adaptive binary and morphology operation,the percentage of black pixels in the human eye area is calculated.Combined with the PERCLOS criterion,the driver's fatigue state is judged by eye closing duration.As the degree of fatigue deepens,the head will appear frequently nodding.On the basis of detecting the change of eye condition,the change of the head posture is detected at the same time to judge the driver's fatigue degree and to alarm.(3)The experiment collects video through a camera.In the thesis,the algorithm is designed and optimized on PC.Build an embedded operating system to detect on the ARM platform and transplant the Qt and OpenCV libraries.At last,Analysis of the results of the experiment.The experiment shows that the system basically achieves the expected design objective.