Driver State Perception Research Based On Digital Image Correlation Method

At present,intelligent automobile has become a research hotspot in the automotive industry,and it is a new industry combining industrialization with information technology.Intelligent driving technology relies on a variety of sensors for environmental information perception to make further judgments,in which driver state perception is indispensable.This paper is mainly based on binocular vision,and has carried out related research on the driver’s driving state perception.Based on binocular vision,a three-dimensional digital image correlation method(3DDIC)and facial key point detection method were used to design a facial motion tracking system,three-dimensional facial reconstruction and facial expression tracking.Tracking the three-dimensional position of the selected point,and then judging the driving state of the driver from the tracked position information,and promptly alerting the driver when the driver is in a fatigue or negative state.An extraction algorithm for the driver’s mouth area is designed.The AdaBoost algorithm is used to detect the face.On the basis of face detection,the facial feature point detection method based on cascade regression is used to detect the key points of the face,and finally the mouth region is extracted.The problem of point tracking and driver facial expression reconstruction based on three-dimensional digital image correlation method is deeply studied,and threedimensional reconstruction and expression tracking based on three-dimensional digital image correlation method are realized.A method to improve the efficiency of 3D-DIC point tracking calculation is proposed,which optimizes and solves the problems in the process of driver facial expression collection and reconstruction.Finally,the face detection,facial key point detection and three-dimensional digital image correlation method are linked together,and the driver’s driving state perception system is designed.The system is verified under the three conditions of yawning,mouth angle rising and mouth corner squatting.The experimental results show that the driver’s state perception system based on digital image correlation method has higher accuracy and can meet the application requirements of intelligent automobiles.