Sistem Deteksi Dan Pengenalan Multi Plat Nomor Kendaraan Dalam Sebuah Citra Tunggal

Lasmana, Pradhitya (2022) Sistem Deteksi Dan Pengenalan Multi Plat Nomor Kendaraan Dalam Sebuah Citra Tunggal. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 07111740000136-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 April 2024. Download (8MB) | Request a copy

Abstract

Penggunaan sistem pengenalan plat nomor kendaraan sudah diaplikasikan di beberapa bidang, seperti dalam sistem parkir di beberapa tempat dan Electronic Traffic Law Enforcement (ETLE). Sistem ini diterapkan guna mendukung program e-Parking di sejumlah instansi pemerintah/swasta dan tilang elektronik (e-Tilang) yang diprakarsai oleh pihak Kepolisian Republik Indonesia (POLRI). Dengan adanya aplikasi sistem pengenalan plat nomor, akan memudahkan petugas parkir dan Polantas agar tidak lagi mencatat nomor plat kendaraan secara manual.
Sebuah sistem pengenalan plat nomor kendaraan telah dirancang dan direalisasikan dalam Tugas Akhir ini menggunakan bahasa pemrograman Python. Sistem yang dibuat memiliki kemampuan mengolah citra digital yang diperlukan dalam beberapa tahapan proses, dimulai dari deteksi lokasi plat nomor dalam citra, segmentasi karakter plat nomor, dan pengenalan karakter dari plat nomor. Berbagai metode telah diaplikasikan untuk mendukung pembuatan sistem tersebut, seperti metode bitwise, thresholding, teknik kontur citra, metode clustering, dan k-nearest neighbor sebagai metode klasifikasinya.
Meskipun berfokus pada deteksi dan pengenalan multi-plat nomor, namun dalam skenario pengujian Tugas Akhir ini juga dilakukan pada sejumlah dataset plat tunggal, dataset warna dasar plat berbeda, dan dataset warna karakter yang berbeda. Dari keseluruhan hasil pengujian terhadap 50 citra plat nomor yang terbagi dalam 5 kelompok dataset, tercatat memiliki keberhasilan deteksi lokasi plat, segmentasi karakter dan pengenalan karakter masing-masing sebesar 98%, 95% dan 87%. Untuk rincian pengujian plat tunggal (30 buah) diperoleh performansi 100% (deteksi lokasi plat), 94% (segmentasi karakter), dan 94% (pengenalan karakter). Sedangkan untuk citra multi-plat nomor (20 buah) memiliki tingkat kesuksesan deteksi sebesar 98%, segmentasi 94.5% dan pengenalan 85%.
==================================================================================================
The use of vehicle number plate recognition systems has been applied in several fields, such as in parking systems in several places and Electronic Traffic Law Enforcement (ETLE). This system is implemented to support the e-Parking program in a number of government/private institutions and electronic ticketing (e-Tilang) initiated by the Indonesian National Police (POLRI). With the application of the number plate recognition system, it will make it easier for parking officers and traffic police officers to no longer record vehicle plate numbers manually.
A vehicle number plate recognition system has been designed and realized in this final project using the Python programming language. The system created has the ability to process digital images that are needed in several stages of the process, starting from the detection of the location of the number plate in the image, segmenting the character of the number plate, and recognizing the character of the number plate. Various methods have been applied to support the creation of the system, such as bitwise method, thresholding, image contour technique, clustering method, and k-nearest neighbor as a classification method.
Although it focuses on detection and recognition of multi-number plates, in this Final Project the test scenario is also carried out on a number of single plate datasets, different base plate color datasets, and different character color datasets. From the overall test results on 50 license plate images divided into 5 dataset groups, it was recorded that they had success in plate location detection, character segmentation and character recognition, respectively 98%, 95% and 87%. For the details of the single plate test (30 pieces) the performance was 100% (plate location detection), 94% (character segmentation), and 94% (character recognition). As for the multi-plate number (20 pieces) the detection success rate is 98%, segmentation is 94.5% and recognition is 85%.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Multi-Plat, Transformasi Wavelet, K-Nearest Neighbor
Subjects:	T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK5102.9 Signal processing. T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK7888.3 Digital computers T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK5101 Telecommunication
Divisions:	Faculty of Electrical Technology > Electrical Engineering > 20201-(S1) Undergraduate Thesis
Depositing User:	Pradhitya Lasmana
Date Deposited:	02 Feb 2022 01:36
Last Modified:	01 Nov 2022 00:57
URI:	http://repository.its.ac.id/id/eprint/92631

Actions (login required)

View Item