Driver Drowsiness Detection System
The Driver Drowsiness Detection System is a real-time computer vision application designed to detect signs of driver fatigue, including drowsiness and yawning, and provide an alert to improve road safety.
Features
- Drowsiness Detection – Detects signs of fatigue using eye movement analysis.
- Yawning Detection – Identifies yawning using mouth movement analysis.
- Facial Landmark Detection – Uses Dlib's 68-point facial landmark model for facial feature detection.
- Eye Aspect Ratio (EAR) – Calculates eye aspect ratio to identify prolonged eye closure.
- Mouth Aspect Ratio (MAR) – Calculates mouth aspect ratio to detect yawning.
- Audio Alert – Triggers an alarm when signs of drowsiness are detected.
- Streamlit Interface – Provides a simple and user-friendly interface for running the application.
- Python
- OpenCV
- Dlib
- Streamlit
- NumPy
- The application captures the driver's video input.
- OpenCV detects the face from the video frames.
- Dlib's 68-point facial landmark model identifies key facial landmarks.
- The Eye Aspect Ratio (EAR) is calculated to monitor eye closure.
- The Mouth Aspect Ratio (MAR) is calculated to identify yawning.
- If the detected values remain beyond the defined thresholds, the system identifies signs of drowsiness.
- An audio alert is triggered to warn the driver.
drowsiness_detection_app.py– Main application code.drowsiness_app.ipynb– Jupyter Notebook containing the development and analysis work.alarm.mp3– Audio alert used by the application.
The objective of this project is to develop a real-time driver monitoring system that can detect early signs of fatigue and provide timely alerts, helping reduce the risk of accidents caused by drowsy driving.
- Improve detection accuracy under different lighting conditions.
- Add real-time driver monitoring and event logging.
- Improve alert mechanisms.
- Explore additional facial and behavioral indicators of fatigue.
Ankita Soni