A Real-time Automated Drowsiness Detection of Driver Using the Visual Behaviour and Machine Learning Technique Through WebCam

M. Santhosh Prakash; K. Ramanan; A. Ananda Kumar; P.M. Sivaraja

Abstract

In our day-to-day life, many accidents can occur due to the drowsiness of the driver. It is one of the causes of road accidents. To prevent accidents, we proposed a system to detect the drowsiness of driver by measuring the eye aspect ratio, mouth yawning and alert the driver. It will save the life of a person. In this, the driver fatigue is continuously captured by webcam. Image processing methods are employed to target the facial region, including the eyes of the driver. The primary objective of the drowsiness detection system is to contribute to the reduction of accidents in both passenger and commercial vehicles. This technology aims to identify initial signs of drowsiness before the driver's alertness significantly diminishes. By doing so, the system will alert the driver to their decreased capability to safely operate the vehicle, thus preventing potential accidents. Our model extracts the face of the driver and detects the eye blinking. If the eye aspect ratio experiences a decrease, the system will signal the driver by emitting a warning sound. The images are acquired by the system using a webcam. After capturing, it detects through Haar Cascade algorithm. Haar Cascade features are additionally utilized to identify the eyes, enabling the assessment of blink frequency. This algorithm aids in discerning whether the eyelids are open or closed. If it is found that the eyes are in closed state, then it detects the driver in drowsy manner and alerts him by the warning sound.

Abstract

Keywords