Evaluasi Penerapan Kinerja Metode Scene Text Recognition Transformer (SVTR) Pada Pengenalan Karakter Plat Nomor Kendaraan Indonesia

Alyani, Nadhila Nurul (2026) Evaluasi Penerapan Kinerja Metode Scene Text Recognition Transformer (SVTR) Pada Pengenalan Karakter Plat Nomor Kendaraan Indonesia. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 2040221049-Undergraduate_Thesis.pdf Restricted to Repository staff only Download (4MB) | Request a copy

Abstract

Peningkatan jumlah kendaraan mendorong kebutuhan akan sistem Automatic License Plate Recognition (ALPR) yang mampu mengidentifikasi jenis kendaraan secara otomatis. Penelitian ini bertujuan untuk mengevaluasi penerapan metode Scene Text Recognition Transformer (SVTR) pada PaddleOCR untuk pengenalan karakter plat nomor kendaraan Indonesia, menganalisis performa YOLOv11 pada tahap deteksi plat nomor, serta mengimplementasikan sistem ALPR berbasis NVIDIA Jetson Orin Nano. Proyek ini menggunakan tiga skenario dataset, yaitu dataset dari Roboflow Plat Nomor Indonesia, Plat nomor crop CCTV dari Dinas Perhubungan, dan gabungan. Performa deteksi dievaluasi menggunakan precision, recall, mAP@50, dan mAP@50–95, sedangkan performa OCR dievaluasi menggunakan Character Accuracy, CER, WER, precision, recall, F1-score, dan Inference Time. Hasil evaluasi menunjukkan bahwa YOLOv11 memperoleh nilai precision hingga 99,93%, recall hingga 98,96%, mAP@50 hingga 99,50%, dan mAP@50–95 hingga 86,19% pada skenario pengujian. Selain itu, model mampu mendeteksi enam jenis plat nomor kendaraan Indonesia dengan tingkat klasifikasi yang baik. Pada tahap pengenalan karakter, SVTR memperoleh Character Accuracy sebesar 97,11%, CER sebesar 2,89%, WER sebesar 21,00%, precision sebesar 88,31%, recall sebesar 88,29%, F1-score sebesar 88,24%, serta Inference Time rata-rata sebesar 41,37 ms. Implementasi sistem ALPR menghasilkan waktu pemrosesan total sebesar 2,54 s mulai dari deteksi plat nomor hingga pengendalian barrier gate.
====================================================================================================================================
The increasing number of vehicles has increased the demand for an Automatic License Plate Recognition (ALPR) system capable of automatically identifying vehicle license plates. This final project aims to evaluate the implementation of the Scene Text Recognition with a Single Visual Model (SVTR) model in PaddleOCR for Indonesian license plate character recognition, analyze the performance of YOLOv11 for license plate detection, and implement an ALPR system for automated parking applications. This project employed three dataset scenarios: the Indonesian License Plate dataset from Roboflow, cropped CCTV license plate images provided by the Department of Transportation, and a combined dataset. The detection performance was evaluated using precision, recall, mAP@50, and mAP@50–95, while the OCR performance was evaluated using Character Accuracy, Character Error Rate (CER), Word Error Rate (WER), precision, recall, F1-score, and Inference Time. The evaluation results show that YOLOv11 achieved up to 99.93% precision, 98.96% recall, 99.50% mAP@50, and 86.19% mAP@50–95 across the evaluation scenarios. In addition, the model successfully classified six types of Indonesian vehicle license plates. For character recognition, the SVTR model achieved a Character Accuracy of 97.11%, a CER of 2.89%, a WER of 21.00%, a precision of 88.31%, a recall of 88.29%, an F1-score of 88.24%, and an average Inference Time of 41.37 ms. The implemented ALPR system achieved a total processing time of 2.54 s, covering the entire process from license plate detection to automatic barrier gate control.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Kata kunci: Automatic License Plate Recognition, YOLOv11, SVTR, EasyOCR, NVIDIA Jetson Orin Nano. Keywords: Automatic License Plate Recognition, YOLOv11, SVTR, EasyOCR, NVIDIA Jetson Orin Nano.
Subjects:	T Technology > T Technology (General) > T57.5 Data Processing T Technology > T Technology (General) > T57.62 Simulation T Technology > T Technology (General) > T58.5 Information technology. IT--Auditing T Technology > T Technology (General) > T58.6 Management information systems T Technology > T Technology (General) > T58.64 Information resources management T Technology > T Technology (General) > T59.7 Human-machine systems.
Divisions:	Faculty of Vocational > 36304-Automation Electronic Engineering
Depositing User:	Nadhila Nurul Alyani
Date Deposited:	13 Aug 2026 00:59
Last Modified:	13 Aug 2026 00:59
URI:	http://repository.its.ac.id/id/eprint/143439

Actions (login required)

View Item