Gunara, Muhammad Adrian (2025) Pengolahan Citra Untuk Deteksi Kantuk Menggunakan Metode Haar Cascade Dan Facial Landmark Detection. Other thesis, Universitas Islam Riau.

Text 203510136.pdf - Submitted Version Restricted to Registered users only Download (3MB) | Request a copy

Abstract

Drowsiness detection is an essential technology for enhancing safety in various activities, such as driving or operating heavy machinery. This study aims to develop an image processing-based drowsiness detection system using the Haar Cascade and Facial Landmark Detection methods. The system captures real-time video from a camera, detects faces using Haar Cascade, and identifies the positions of the eyes and mouth through Facial Landmark Detection. The primary parameters used are the Eye Aspect Ratio (EAR) and Mouth Aspect Ratio (MAR), which are calculated to analyze eye openness and mouth states. When the EAR value falls below a certain threshold for several frames, the system detects drowsiness and provides a warning. Similarly, if the MAR value exceeds a certain threshold for several frames, the system detects yawning and triggers an alarm. The model was developed using the Support Vector Machine (SVM) algorithm, achieving an accuracy of 75% on the test dataset. The results demonstrate that the system can detect drowsiness in realtime with high precision, making it suitable for implementation in various safety applications.

Actions (login required)

View Item