Pengembangan Sistem License Plate Recognition Menggunakan YOLO dan OCR pada Gerbang ITS

Ghifari, Alvian (2024) Pengembangan Sistem License Plate Recognition Menggunakan YOLO dan OCR pada Gerbang ITS. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5027201035-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 April 2026. Download (2MB) | Request a copy

Abstract

Dengan 148 juta kendaraan di Indonesia, pencurian kendaraan roda dua mencapai 13.607 kasus pada 2022. Kejadian serupa terjadi di tempat umum, termasuk kampus seperti Institut Teknologi Sepuluh Nopember (ITS). Sistem gerbang tradisional di ITS, sering menghadapi antrian panjang di pintu keluar karena pengecekan STNK yang dilakukan secara manual. Sistem ini juga berpotensi mengakibatkan hilangnya STNK milik pemilik kendaraan. Oleh karena itu, penerapan License Plate Recognition (LPR) atau sistem pengenalan pelat nomor kendaraan dengan menggunakan teknologi semakin relevan. Implementasi LPR pada sistem gerbang yang ada menjadi penting untuk mengatasi masalah tersebut dan meningkatkan efektivitas keamanan di ITS. Penelitian ini bertujuan mengembangkan sistem gerbang masuk dan keluar otomatis yang efisien dan akurat dalam mengenali nomor registrasi pelat kendaraan di ITS. Solusi yang diperkenalkan dalam penelitian ini adalah penerapan teknologi LPR yang terintegrasi dengan gerbang keluar-masuk kendaraan di ITS. Dalam implementasinya, dibandingkan dua metode algoritma YOLO yaitu deteksi objek dan segmentasi untuk mendeteksi pelat nomor, yang kemudian menghasilkan koordinat lokasi pelat nomor. Koordinat ini digunakan dalam proses pre-processing sebelum dijadikan input dalam proses pengenalan karakter menggunakan OCR. Service LPR ini mampu mengidentifikasi pelat nomor dengan waktu 1 detik. Akurasi kemiripan pelat nomor yang dibaca memiliki rata-rata 83,38 persen untuk pelat nomor berwarna hitam dan 75,74 persen untuk pelat nomor berwarna putih.
===================================================================================================================
With 148 million vehicles in Indonesia, motorcycle theft reached 13,607 cases in 2022. Similar incidents occur in public places, including campuses such as the Institut Teknologi Sepuluh Nopember (ITS). At ITS, the traditional gate system often faces long queues at the exit due to manual checking of vehicle registration certificates (STNK). This system also has the potential to result in the loss of STNK documents belonging to vehicle owners. Therefore, the implementation of License Plate Recognition (LPR) technology for vehicle plate number recognition becomes increasingly relevant. The integration of LPR into the existing gate system is crucial to address these issues and enhance security effectiveness at ITS. This research aims to develop an efficient and accurate automated entry and exit gate system capable of recognizing vehicle registration numbers within ITS. The proposed solution in this study involves the application of LPR technology integrated with the vehicle entry and exit gates at ITS. In its implementation, two YOLO algorithm methods, object detection, and segmentation, are compared to detect license plates, subsequently yielding the coordinates of the plate's location. These coordinates are utilized in a pre-processing stage before being input into the character recognition process using OCR. The LPR service can identify license plates within 1 second. The average similarity accuracy of the recognized license plates is 83.38% for black license plates and 75.74% for white license plates.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	OCR, License Plate Recognition, YOLO, Gate, Gerbang, Pelat Nomor
Subjects:	T Technology > T Technology (General)
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Information Technology > 59201-(S1) Undergraduate Thesis
Depositing User:	Alvian Ghifari
Date Deposited:	05 Feb 2024 03:24
Last Modified:	05 Feb 2024 03:24
URI:	http://repository.its.ac.id/id/eprint/106049

Actions (login required)

View Item