Implementation of Motorist Weariness Detection System using a Conventional Object Recognition Technique

Detecting driver drowsiness is a huge crucial problem in the sector of accident-avoidance technologies, so the development of an innovative intelligent system came into the picture. The system also prioritized safety concerns such as informing the victim and avoiding yawning. The technique for this system is a machine learning-based sophisticated algorithm that can identify the driver's facial expressions and quantify the rate of driver sleepiness. This may be avoided by activating an alarm that causes the driver to become alert when he or she becomes fatigued. The Eye Aspects Ratio (EAR) is used to recognize the system’s drowsiness rate by calculating the facial plot localization which extracts and gives the drowsiness rate.Current approaches, however, have significant shortcomings due to the considerable unpredictability of surrounding conditions. Poor lighting may impair the camera's ability to precisely measure the driver's face and eye. This will affect image processing analysis which corresponds to late detection or no detection, tendering the technique in accuracy and efficiency. Numerous strategies were investigated and analyzed to determine the optimal technique with the maximum accuracy for detecting driver tiredness. In this paper, the implementation of a real-time system is proposed that requires a camera to automatically trace and process the victim’s eye using Dlib Python, and OpenCV. The driver's eye area is continually monitored and computed to assess drowsiness before generating an output alarm to notify the driver.