Ardian, Mohammad Faras (2023) Analisis Ekspresi Wajah Pengemudi Mobil Untuk Deteksi Kantuk Secara Real-Time Menggunakan Metode YOLOV5. Undergraduate thesis, UPN Veteran Jawa Timur.
|
Text (COVER)
19082010045-cover.pdf Download (1MB) | Preview |
|
|
Text (BAB 1)
19082010045-bab1.pdf Download (86kB) | Preview |
|
![[img]](http://repository.upnjatim.ac.id/style/images/fileicons/text.png) |
Text (BAB 2)
19082010045-bab2.pdf Restricted to Registered users only until 4 June 2025. Download (635kB) | Request a copy |
|
|
Text (BAB 3)
19082010045-bab3.pdf Restricted to Registered users only until 4 June 2025. Download (268kB) | Request a copy |
|
|
Text (BAB 4)
19082010045-bab4.pdf Restricted to Registered users only until 4 June 2025. Download (1MB) | Request a copy |
|
|
Text (BAB 5)
19082010045-bab5.pdf Download (19kB) | Preview |
|
|
Text (Daftar Pustaka)
19082010045-daftarpustaka.pdf Download (145kB) | Preview |
Abstract
Traffic accidents caused by driver drowsiness or fatigue are a serious problem in road transportation. This study proposes the use of a drowsiness detection application that uses the YOLOv5 method to detect signs of sleepiness on the driver's face and tests the accuracy and performance of the drowsiness detection application based on factors such as lighting level, user distance, user characteristics and delay time. The research method used in this study is the AI Project Cycle method. The dataset is obtained through the Roboflow platform. This research involves the development of the YOLOv5 algorithm to detect signs of drowsiness such as closed eyes or tilted head of the driver through the front camera of a smartphone device. Furthermore, model evaluation is carried out using a confusion matrix and a precision-recall curve. In addition, the drowsiness detection application can provide a warning with an alarm when the driver is detected as drowsy and alert notifications based on the history of previous events. The results of this study indicate that a car driver's drowsiness detection system using the YOLOv5 Algorithm has been successfully developed and installed on a smartphone with high accuracy. The model produces a good level of accuracy, precision and recall, namely 95%, 94% and 96%. The drowsiness detection application is also able to provide a warning to drivers who are detected as drowsy with a threshold > 0.4. In addition, the ISO test best detects the driver's condition during the day with ISO Lux > 1000, measuring distance of 150 degrees and delay time of 2.4 seconds. Keywords: Traffic Accident, Driver's Drowsiness, YOLOV5 Algorithm, Drowsiness Detection Application, Roboflow
| Item Type: | Thesis (Undergraduate) | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Contributors: |
|
||||||||||||
| Subjects: | T Technology > T Technology (General) > T58.6-58.62 Management Information Systems | ||||||||||||
| Divisions: | Faculty of Computer Science > Departemen of Information Systems | ||||||||||||
| Depositing User: | Mohammad Faras Ardian | ||||||||||||
| Date Deposited: | 05 Jun 2023 03:15 | ||||||||||||
| Last Modified: | 05 Jun 2023 03:15 | ||||||||||||
| URI: | http://repository.upnjatim.ac.id/id/eprint/14435 |
Actions (login required)
 |
View Item |