An Automatic Early Alert System on Detecting Dozing Driver

Abstract

Driver fatigue due to continuous work or improper sleep is one among the foremost reasons for lethal road accidents across the globe. In order to avoid accidents, driver condition needs to be monitored and an alert signal should be given to avoid any accidents. The existing solutions to detect drowsiness involve expensive sensors and costly devices. In this work, a real time, light weight and less costly framework and implementation for detecting driver’s drowsiness is proposed. This paper proposes an effective drowsiness detection system for driver using eye movement and yawning detection. The collective use of mouth and eye condition detection, i.e., if yawning is detected with eyes closed, fetches better information about the detection of fatigue or drowsiness in driver. In order to evade any critical situation, the proposed work will give a sound alert signal to alert the driver and suggest switching the transmission mode from manual to autopilot. The system detects doziness of driver by detecting yawning and closing of eyes as a precautionary measure by using face images captured through camera. The experimental result shows that the suggested approach being real time and lightweight also performs well.