Driver drowsiness detection system using artificial vision

Abstract

This document presents the development of a prototype to detect a state of drowsiness in drivers, for which the artificial vision technique is used based on the association of the state that users have when blinking for a long time, thus detecting possible drowsiness. The solution consists of a portable prototype designed with microcomputers due to its easy transport and its adaptability to facial recognition programs. The design includes a Raspberry PI module and OpenCV facial recognition libraries programming with the Python programming language. This device is adaptable to different drivers, taking a video sequence to generate a drowsiness state pattern in drivers, which, if it exceeds a threshold of $80 \%$, will emit an audible and a visual alarm, which will put the user in sleep mode, alert, preventing any accident.