Driver Drowsiness Detection

Abstract

The state of the driver of being extremely tired or sleepy through the operation of the vehicle is called driver drowsiness. Different factors caused this state such as alcohol, lack of sleep, and the side effect of some medication. The drowsiness of drivers is a serious safety lead to accidents or fatalities on external and internal roads. The increased number of road accidents resulted from drowsy driving. A special smart, reliable, and accurate system, Using Python language 3.6 for Windows, was designed to build an alert system for drivers in detecting drowsiness driver. This system is crucial in reducing accidents road by the ability to concentrate, react quickly, and produce sound decisions through driving. This system implements a real-time detector that can monitor the states of drivers through driving.
Smart cameras with 16-megapixel were used to ensure that capturing photos have a high quality. These cameras were used in gathering the driver’s dataset in different alertness states, including both alert states and drowsy. The collected dataset is processed by extracting all relevant features such as head movement, yawning, and eye closure, which were used in identifying the driver’s drowsiness. Python’s libraries such as TensorFlow, OpenCV, Keras, and Pygame are used for extracting all the above features. Viola-Jones algorithm is used in face eye region detecting and extracting from the image of the face in the proposed system. A Support Vector Machine (SVM) algorithm was used in classifying between drowsy and non-drowsy drivers. The system is tested and evaluated in the real world, to ensure that the system is reliable and robust; it has high performance and accuracy, and the accuracy is about 99.1%. This system can be used in manufacturing vehicles.