Fatigue detection system in cars
Fatigue among drivers continues to be one of the leading causes of road accidents particularly when people are driving for long hours, during the night, or in conditions that demand high focus. Even though technology has come a long way, a lot of cars still don’t come with proper systems that can de...
Saved in:
| Main Author: | |
|---|---|
| Format: | Final Year Project / Dissertation / Thesis |
| Published: |
2025
|
| Subjects: | |
| Online Access: | http://eprints.utar.edu.my/6978/1/fyp_CT_2025_VK.pdf http://eprints.utar.edu.my/6978/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Fatigue among drivers continues to be one of the leading causes of road accidents particularly when people are driving for long hours, during the night, or in conditions that demand high focus. Even though technology has come a long way, a lot of cars still don’t come with proper systems that can detect when the driver is starting to lose focus or show early signs of drowsiness. That is what motivated this project – to come up with a solution that can identify fatigue symptoms through facial cues in real time and alert the driver before things get dangerous.
The project uses a lightweight CNN called Mobilenet to perform real-time image-based detection efficiently on a compact device like the Raspberry Pi. It integrates EAR and MAR to detect eye closure and yawning, while also tracking blink frequency, which can help identify early signs of drowsiness, Facial landmarks are used to extract these features from the driver’s face using a webcam. To further enhance reliability, the system includes head tilt detection by calculating pitch and roll angles to recognize unnatural head positions commonly associated with fatigue.
Special attention is given to real-world challenges such as low lighting conditions, reflection from spectacles, and obstructions that may affect facial visibility. These factors can impact detection accuracy, so the system is designed to be robust and adaptable in various driving environments.
Overall, this project is more than deep learning experiment. It integrates computer vision, facial analysis, real-time processing, and system integration to deliver a lightweight and practical solution. The system is cost-effective, easy to deploy, and reliable enough to help reduce the risk of accidents caused by driver fatigue.
|
|---|