Research On Identification Methods And Application Of Fatigue Driving Eye Feature

Currently, fatigue driving has become one of the main reasons of frequent traffic accidents,to people’s lives and property caused huge losses, so, if you can advance the state of the driver’sfatigue is detected and gives early warning, can effectively avoid accidents caused by driverfatigue.Fatigue can cause the body’s physiological characteristics change, and changes in thehuman eye is particularly prominent in the human eye fatigue duration of closure will grow.Based on the principle, this study is characterized by fatigue driving everyone eye recognitionand applied research.This paper first of all domestic and foreign related technology of fatigue driving detectionwere analyzed, the gray integral projection algorithm and Kalman filter the most simple,effective tracking algorithm to detect fatigue. The primary task of fatigue detection is facedetection and eyes localization, detection and location before the first image preprocessing,including similarity calculation, gray processing and two value processing, image pretreatmentcan reduce the influence of illumination and noise, and then use the gray integral projectionalgorithm of face detection and eyes localization. After the completion of the human eyepositioning Kalman filter algorithm eye tracking, eye tracking is designed to save time, in orderto meet the needs of real-time fatigue detection system, standardized operation and then theimage of the human eye can be avoided because affect the driver’s posture detection result.Finally, the fatigue state judgment, firstly by the human eye contour extraction, amplificationand eye area calculation sheets to determine the degree of open-eye or closed state of thehuman eye, the paper sheets of the human eye is less than20%of the degree of eye imagesconsidered closed, according PERCLOS criteria to determine the driver’s fatigue. Fatiguetesting process, the accuracy to the guarantor eye tracking, in a unit time required to implementrelocation to the human eye, to prevent fatigue caused by the failure detecting tracking errors.In this paper, the DaVinci experimental system as a platform, use a PC combined with theintegrated development environment CCS simulation fatigue testing, which tests including thedriver were in a different light, different attitude and a different frame fatigue detection, toachieve the expected results goals. According to the experimental results analysis showscompleted, under normal circumstances, the system real-time and accurate face detection, eyelocation and tracking, fatigue judgment also shows that the system is suitable for the driverfatigue detection.