Design of Fatigue Driving Behavior Detection Based on Circle Hough Transform

Area of Science:

• Occupational Health
• Human-Computer Interaction
• Biomedical Engineering

Background:

• Chronic fatigue is a major risk factor for occupational and traffic accidents.
• Fatigue driving significantly impairs traffic quality, flow efficiency, and incurs societal costs.
• Existing fatigue detection methods may be invasive or lack systematic accuracy.

Purpose of the Study:

• To propose and evaluate a noninvasive machine vision-based method for detecting driver fatigue.
• To develop a system that accurately identifies fatigue indicators from facial and head movement data.
• To implement an intelligent auxiliary device for real-time fatigue alerts to high-risk drivers.

Main Methods:

• Utilized a machine vision algorithm with optimized skin color segmentation and edge detection.
• Implemented eye contour extraction and an optimized Hough transform algorithm.
• Analyzed facial features including mouth aspect ratio (MAR), eye condition (PERCLOS-80), and head nodding frequency.

Main Results:

• The developed system accurately and systematically detects fatigue indicators.
• An ideal eye threshold of 0.3, PERCLOS-80 standard, MAR, and head nod frequency were identified as key metrics.
• The method demonstrated effectiveness in identifying fatigue in facial image data analysis.

Conclusions:

• The proposed noninvasive machine vision method offers an accurate and systematic approach to fatigue detection.
• This technology can effectively alert high-risk drivers, preventing involuntary engagement in dangerous activities.
• The system has the potential to significantly improve road safety by mitigating fatigue-related accidents.