Driving State Of The Computer Vision-based Driver Safety Monitoring System

Driving fatigue is one of the most important factors in traffic accident and many attentions were paid on fatigue preventing and alarming during the past years. Fatigue alarming includes fatigue detection, evaluation and vaillance system development.In this paper, advances on the research of the driving fatigue are studied, and PERCLOS is selected to evaluate driving fatigue, then a computer vision based driving- fatigue-monitor system is developed.The method on eyes location based on a skin model and the face's geometry model is studied, then it is modified considering real requirement. Experimental results show that the modified method performs better, especially when eyes close.After eyes location, eyes area is further processed and PERCLOS is chosen to be the criterion of the driving fatigue.Experimental results show that our algorithm performs well in reliability and the real-time performance.