Fatigue Driving Detection Based on Machine Learning and Image Processing Technology

Abstract

Fatigue driving is a major cause of traffic accidents. If fatigue driving can be detected and stopped in time, perhaps traffic accidents will be reduced. This paper proposes a fatigue driving detection algorithm using machine learning and image processing technology. The algorithm is designed to run on the vehicle’s smart terminal. In the algorithm, terminal has to analyze the status of driver’s eyes, nose and mouth using machine learning and image processing technology. When catch typical behaviors such as blink, yawn and nod, terminal will record the time and calculate the frequencies of these behaviors. If one of the frequencies is greater than the threshold, terminal will trigger the alert and take measures to make driver no longer fatigue driving. The algorithm works fine in the experiment. The experimental accuracy of the algorithm is 94%.