Claim Missing Document
Check
Articles

Found 1 Documents
Search

Deteksi Kantuk pada Pengemudi melalui Jumlah Kedipan Mata Menggunakan Facial Landmark berbasis Intel NUC Dewi Amalia; Fitri Utaminingrum
Jurnal Pengembangan Teknologi Informasi dan Ilmu Komputer Vol 5 No 12 (2021): Desember 2021
Publisher : Fakultas Ilmu Komputer (FILKOM), Universitas Brawijaya

Show Abstract | Download Original | Original Source | Check in Google Scholar

Abstract

The level of traffic accidents is increasing every year, one of the causes is the driver who is tired and sleepy when driving. Therefore a system will be made to anticipate accidents by giving warnings in the form of writing and alarm. The system uses Intel NUC for processing, the webcam accepts inputs and monitors to see image captured by the camera and see information on the condition of the eyes and sleepy information. The method used is Facial Landmark for detection of eye areas on the face. For under-lighting or uneven lighting, use the feature of image thresholding, namely adaptive threshold gaussian. Detection of the eye area on the face and sleepiness is done with a camera within 30cm, 40cm and 50cm parallel to the shoulder or chest. This detection also uses the range of light intensity 0-49 lux and 50-400 lux. The average accuracy of Facial Landmark for detecting eye areas on the face with light intensity 0-49 lux is 93.33% and for light intensity 50-400 lux is 100%. While the average accuracy of drowsiness detection at 0-49 lux light intensity is 96.66% and for light intensity 50-400 lux is 98.88%. The average system accuracy is 97.77%. The fastest computing time of the system at 0-49 lux light intensity is 0.33 s and at light intensity 50-400 lux that is 0.34 s.