This research addresses the escalating issue of motor vehicle accidents in Malaysia, specifically focusing on the critical challenge of driver fatigue. The study highlights the prevalence of road accidents attributed to tiredness, emphasizing the need for effective detection and prevention methods. Current sleepiness detection technologies are critiqued for their expense and impracticality during driving. The primary objective of this research is to develop a driving fatigue detection system using the Haar Cascade technique, specifically analyzing the closure of the driver's eyes. Additionally, the research aims to design an emergency notification system triggered by the detection of driver fatigue and to evaluate the system's performance. The study employs a Python-based platform and utilizes Raspberry Pi 4 Model B, Raspberry Pi Camera, and Buzzer for hardware implementation. The flow chart outlines the sequential steps, from image acquisition to the activation of the buzzer and notification system. The results indicate prompt and accurate eye closure detection, with an average response time of 0.4 milliseconds. However, limitations, particularly false detections involving spectacles, are identified, yielding an overall accuracy of 83%. The future work suggests incorporating machine learning techniques, adjusting Haar Cascade parameters based on real-time factors, and integrating supplementary sensors for a more comprehensive fatigue detection approach. Collaboration with automotive manufacturers, continuous system calibration, and user-friendly interfaces are proposed for the advancement and wider adoption of the driving fatigue detection system
Davanathan et al. (Tue,) studied this question.