Research on Early Warning of Driver Fatigue Status Based on Image Processing

As a major factor of traffic safety, driving fatigue has a great impact on life and property. A non-contact, real-time and high-precision fatigue detection method is studied in this paper by using the bi-directional projection integral for eye positioning based on the characteristics of the different reflectivity of light in different areas of the face, utilizing the PERCLOS method to detect the frequency of the driver's eye opening and closing according to the change curve of pupil radius, so ad to determine whether it is in a fatigued driving state. The accuracy of the method has been approved by the simulation results, therefore, the method can be applied to fatigue detection effectively.