Research on Driving Fatigue Detection Based on PERCLOS

Abstract

This paper expounded the mechanism on PERCLOS detecting and evaluating fatigue driving, and briefly introduced the composition of hardware evaluation system. This system firstly detected the face region roughly using skin-color model. Then the drivers’ eyes were located exactly on RGB by face geometry features. Finally, a judgment of the drive’s fatigue condition was made according to the number of white pixels in the eyes area and the duration of driver’s eyes closing. Experiments showed that the system could accurately locate eyes and the precision rate was 93.3%.