ABSTRACT:
Driver drowsiness is a major cause of road accidents worldwide, leading to severe injuries and loss of life. Continuous monitoring of driver alertness is essential to reduce such incidents. This project presents a real-time Driver Drowsiness Detection System that uses computer vision and embedded systems to identify signs of fatigue and alert the driver effectively.
The system is implemented using a Raspberry Pi integrated with a camera module for real-time video acquisition. A deep learning-based object detection model, YOLO (You Only Look Once), is used to detect the presence of a person (driver) in the video frame. Once the driver is detected, the region of interest is extracted and processed further for facial analysis.
For detailed facial feature analysis, MediaPipe Face Mesh is employed to detect facial landmarks with high accuracy. Using these landmarks, key indicators of drowsiness such as Eye Aspect Ratio (EAR) and Mouth Aspect Ratio (MAR) are calculated. EAR is used to monitor eye closure over consecutive frames, while MAR is used to detect yawning behavior. If the system observes prolonged eye closure or frequent yawning, it classifies the driver as being in a drowsy state.
Upon detecting drowsiness, the system activates alert mechanisms to ensure driver safety. A warning message, “DROWSINESS DETECTED,” is displayed on a 16x2 LCD screen, and an audible alert is generated using a buzzer to immediately warn the driver. When the driver is alert, the system displays a normal status message indicating safe driving conditions.
The integration of YOLO for person detection and MediaPipe for facial landmark analysis ensures both accuracy and robustness in real-time conditions. The system is designed to be low-cost, portable, and efficient, making it suitable for deployment in vehicles. This project demonstrates the practical application of artificial intelligence, computer vision, and embedded systems in enhancing road safety and preventing accidents caused by driver fatigue.

INTRODUCTION:
Road accidents caused by driver fatigue and drowsiness have become a serious global concern, contributing significantly to injuries and fatalities. Long driving hours, lack of sleep, and monotonous road conditions often lead to reduced alertness, slower reaction times, and impaired decision-making abilities. Therefore, developing an efficient system to monitor driver alertness in real-time is essential for improving road safety.
With the advancement of artificial intelligence and computer vision technologies, it is now possible to design intelligent systems that can analyze human behavior and detect signs of fatigue. This project focuses on developing a Driver Drowsiness Detection System using a Raspberry Pi, which continuously monitors the driver through a camera and analyzes facial features to determine the level of alertness.
The system utilizes YOLO (You Only Look Once), a deep learning-based object detection algorithm, to identify the presence of a driver in the video frame. Once the driver is detected, the facial region is processed using MediaPipe Face Mesh, which provides detailed facial landmarks. These landmarks are used to compute parameters such as Eye Aspect Ratio (EAR) and Mouth Aspect Ratio (MAR), which are reliable indicators of eye closure and yawning, respectively.
By analyzing these parameters over time, the system can accurately detect drowsiness conditions. When drowsiness is detected, the system immediately alerts the driver through a visual message displayed on a 16x2 LCD and an audible warning using a buzzer. This real-time alert mechanism helps prevent accidents by prompting the driver to regain alertness.
The proposed system is designed to be cost-effective, portable, and suitable for real-world applications. By combining embedded systems with advanced computer vision techniques, this project aims to contribute to the development of safer transportation systems and reduce accidents caused by driver fatigue.


OBJECTIVES:
The primary objective of this project is to develop an intelligent and real-time Driver Drowsiness Detection System that can continuously monitor the driver’s condition and provide timely alerts to prevent accidents caused by fatigue. The system aims to utilize advanced computer vision techniques along with embedded system capabilities to analyze the driver’s facial behavior in real time. By integrating deep learning algorithms such as YOLO for person detection and MediaPipe for facial landmark analysis, the system ensures accurate detection of drowsiness indicators like eye closure and yawning. Furthermore, the project focuses on designing an efficient, low-cost, and portable solution that can be implemented in real-world vehicle environments to enhance overall road safety.
• To develop a real-time video monitoring system using a Raspberry Pi and camera module.
• To implement YOLO (You Only Look Once) algorithm for detecting the presence of the driver in the frame.
• To utilize MediaPipe Face Mesh for extracting facial landmarks and analyzing facial features.
• To calculate Eye Aspect Ratio (EAR) and Mouth Aspect Ratio (MAR) for detecting eye closure and yawning.
• To design an alert mechanism that displays warning messages on a 16x2 LCD and activates a buzzer when drowsiness is detected.
• To optimize the system for efficient performance on embedded hardware like Raspberry Pi.
• To develop a cost-effective and portable solution suitable for real-time vehicle applications.
• To improve road safety by reducing accidents caused due to driver fatigue and drowsiness.