Driver Drowsiness and Distraction Detection System

Problem statement

Drowsy and distracted driving are major causes of road accidents. Human drivers often fail to recognize their own fatigue levels. A system that can detect early signs of drowsiness and issue warnings could prevent fatal accidents.

Abstract

This project develops a real-time driver monitoring system using facial landmark detection and gaze estimation. It tracks blink rate, eye closure duration, head nodding, and gaze direction. If the system detects prolonged eye closure or off-road gaze, it triggers audible and visual alerts. A CNN-based classifier can improve accuracy under different lighting conditions.

Components required

• Camera module / webcam
• OpenCV + Mediapipe or Dlib
• Python/TensorFlow
• Buzzer / speaker for warnings
• Dashboard (optional)

Block diagram

Working

The camera continuously captures frames of the driver's face. Eye aspect ratio and blink rate are computed from facial landmarks. If eyes remain closed beyond a threshold, the system triggers a drowsiness alert. Gaze tracking checks whether the driver is looking away from the road. Head tilt detection identifies nodding associated with sleepiness. Alerts include sound, vibration, or dashboard indicators.

Applications

• Smart vehicles
• Fleet management
• Driver safety monitoring
• Semi-autonomous driving research