Driver Drowsy Detection System Using Machine Learning and Open CV

Driver fatigue is another significant risk factor that claim a fairly good proportion of the accident rates globally. Fatigue affects the important human functions including decision making, ability to concentrate and reaction time making it of paramount importance to have good counter measures in place. That is why, the proposed Driver Drowsiness Detection System in this paper focuses on real-time video surveillance and machine learning to identify symptoms of drowsiness. The drowsiness detection of the system is made possible by integrating OpenCV and other computer vision algorithms that take pictures of the face via a camera and examine for patterns that are characteristic of drowsiness, which include yawning, blinking of the eyes frequently, or the closing of the eyes for too long. That is, the use of OpenCV in facial landmark detection and other computations such as EAR and MAR with the help of two well trained models on deep learning. The system uses adaptive thresholds and sophisticated preprocessing to ensure its performance stays high regardless of the conditions which include light, camera angle, and facial expressions. Such consistencies of both the methodologies in convergence help the system to assess the driver's alertness level and immediately produce an alert the moment the driver looks sleepy. Intended to be deployed in semi-autonomous vehicles, vehicles in commercial fleets, as well as individual automobiles, the proposed system can be implemented at a large-scale to help increase road safety. Thus, eliminating the risks related to fatigue-related accidents, it opens an opportunity to minimize fatality rates and change the future environment in transport. The technical issues, the difficulties and the realistic uses of the system are also described This paper focuses on the contribution of the system to enhance road safety and reduce the number of deaths.