Aji, Geo Ardana Ihsan Purnama (2026) Pengembangan Sistem Deteksi Kendaraan Berbasis Yolov8 Untuk Pengenalan Akses Masuk Penghuni Di Kawasan Perumahan. Undergraduate thesis, Universitas Hayam Wuruk Perbanas.
![[img]](http://eprints.perbanas.ac.id/style/images/fileicons/text.png) |
Text
ARTIKEL ILMIAH.pdf Restricted to Registered users only Download (887kB) |
|
|
Text
COVER.pdf Restricted to Registered users only Download (972kB) |
|
|
Text
BAB I.pdf Download (244kB) | Preview |
|
|
Text
BAB II.pdf Download (653kB) | Preview |
|
|
Text
BAB III.pdf Restricted to Registered users only Download (1MB) |
|
|
Text
BAB IV.pdf Restricted to Registered users only Download (2MB) |
|
|
Text
BAB V.pdf Download (224kB) | Preview |
|
|
Text
LAMPIRAN.pdf Restricted to Registered users only Download (3MB) |
Abstract
This study aims to develop a YOLOv8-based vehicle detection system to automatically identify resident and non-resident vehicle access in residential areas. The system contributes to the application of computer vision and artificial intelligence in residential security by integrating vehicle type detection, license plate recognition, a web-based information system, and real-time notifications. The main idea of this research is to utilize YOLOv8 to detect vehicle objects and license plates through CCTV cameras, then apply EasyOCR to read Indonesian license plate characters, which are subsequently stored in a database and displayed through an information system. The research adopts the Waterfall method, covering requirement analysis, dataset collection, model training, system implementation using Python, OpenCV, EasyOCR, and Flask, as well as system testing using black box testing and model performance evaluation. The test results show that the model achieves a Precision of 97.9% and a Recall of 96.3%, with an F1-Score of 97.1%, indicating a strong balance between detection accuracy and completeness. In addition, the mAP@50 reaches 98% and the mAP@50–95 reaches 85.1%, demonstrating consistent detection performance across different Intersection over Union levels. The EasyOCR accuracy evaluation uses Character Accuracy with an experiment of reading 50 license plates, all of which are correctly recognized, resulting in a Character Accuracy of 100%. All features of the information system function according to user requirements. The implications of this study include improved efficiency and accuracy in vehicle access recording, reduced reliance on manual processes, and the potential application of the system as a digital security model for other residential areas.
| Item Type: | Thesis (Undergraduate) |
|---|---|
| Subjects: | 600 - TECHNOLOGY > 600 - 609 TECHNOLOGY > 600 - TECHNOLOGY |
| Divisions: | Bachelor of Informatics |
| Depositing User: | GEO ARDANA IHSAN PURNAMA AJI |
| Date Deposited: | 27 Mar 2026 04:29 |
| Last Modified: | 27 Mar 2026 04:29 |
| URI: | http://eprints.perbanas.ac.id/id/eprint/14007 |
Actions (login required)
 |
View Item |