DETECTION OF DRIVERS’ DROWSINESS

Abstract

Making driver drowsiness detection in embedded systems meet real-time requirements is a challenging problem; in the meantime, some issues remain unresolved, such as drivers' heads tilting and inadequately sized eye images. The goal of this project is to develop an effective approach for identifying the eye states of drivers and detecting their drowsiness in embedded systems using image processing techniques. This technique uses face and eye detection to initialize the location of the driver's eyes, breaking the conventional approach to drowsiness detection and making it real-time. Subsequently, an object tracking technique is employed to monitor the eyes; ultimately, the identified eye state allows us to determine the driver's drowsiness state with PERCLOS. The outcomes of the experiment indicate that it makes good agreement with analysis. Keywords: Drowsiness detection, Face detection, Eye location, Object tracking.