The Design Of The Car Driver Fatigue Protection Instrument

Fatigue driving has become the main cause of the traffic accidents. According tothe survey,20%of the traffic accident is caused by fatigue driving.This problem hasattracted widespread attention. In order to solve this problem, the paper designs DriverFatigue Detection System, which detects drivers’ eyes state mainly through the facedetection technology and the human eye test recognition technology, and then judgeswhether drivers are in fatigue and gives the alarm.Eyes location and identification of eyes state is the key of the system, for thecharacteristics and state of eyes not only can describe the facial features, but also candescribe one’s mind state. Only completing the eyes location and recognition of eyesstate can achieve judgment of driver fatigue. However, there are massive differencesamong the size, color and shape of eyes; the dynamic characteristics of eyes andbackground light and some other reasons bring great difficulties for eyes testing. Thispaper proposes an eyes-positioning method based on texture segmentation. The methodapplies the texture segmentation to detect and extract the face region first, and thenmakes use of facial features distribution and its texture segmentation to realize accuratelocation of eyes. After completing eyes location, combining Kalman filtering andtemplate matching to realize the eye tracking, and to identify the state of eyes throughtemplate matching. Finally PERCLOS is adopted to judge whether the driver is infatigue.