Real-time Fatigue Detection Based On Dlib Model And Head PostureEstimation

Fatigue phenomenon appears in a variety of occasions in daily life,in the driving environment for fatigue detection is particularly important.Along with rising of living standards,cars have entered the ordinary people,has become an important means of transportation,People’s Daily travel traffic accident prevention problem has became the hot topic of social and scientific research,including trafic rules of updating,the urban eye system and strictly control the traffic department for speeding,overloading and ease of drunken driving accident,but fatigue driving accident stil can not get effective solution.Therefore,fatigue detection in the fatigue driving problem has become a national concern,but also become a guarantee of people’s life and property safety and reduce traffic accidents practical problems.Based on the research at home and abroad of various applied in fatigue test method,this paper proposes a model based on dlib and Head Pose Estimation method of real-time fatigue detection,Uses the camera real-time acquisition head video data,using OpenCV face detection,through the dlib model can extract 68 organs in face feature points,the first to use the EAR algorithm and MAR algorithm can distinguish between the behavior of the eyes and mouth,setting up reasonable threshold according to the experiment to determine fatigue behavior,reuse of head pose estimation method,according to the pitch Angle and roll Angle information comparehensive judgment fatigue state.In this paper,face detection,feature point extraction and fatigue detection are verified.The fatigue detection conditions obtained through the experiment are tested for drivers in the driving environment,and the fatigue warning message is sent out when the fatigue conditions are reached.he experimental results show that the system has a high accuracy for fatigue detection,can meet the fatigue detection needs,and achieve the fatigue detection warning effect.