Deteksi Plat Nomor Kendaraan Bermotor Menggunakan Metode Sliding Concentric Windows (SCW) untuk Aplikasi Sistem Transportasi Cerdas

Firswandy, Rangga Imantaka (2019) Deteksi Plat Nomor Kendaraan Bermotor Menggunakan Metode Sliding Concentric Windows (SCW) untuk Aplikasi Sistem Transportasi Cerdas. Undergraduate thesis, Institut Teknologi Sepuluh Nopember.

[thumbnail of 07111240000188-Undergraduate_Thesis.pdf]
Preview
Text
07111240000188-Undergraduate_Thesis.pdf

Download (2MB) | Preview

Abstract

Sistem transportasi cerdas semakin menjadi kebutuhan saat ini, sehingga berbagai teknik dan algoritma License Plate-Recognition (LPR) terus dikembangkan untuk aplikasi pengenalan plat nomor kendaraan bermotor. Namun, berbagai kendala masih banyak ditemui dikarenakan kondisi plat nomor yang tidak standar (warna, bentuk, ukuran dan pola), faktor cuaca, level pencahayaan, dan resolusi kamera pindai turut menentukan pengaruh kualitas citra yang akan dideteksi.
Fokus dari tugas akhir ini adalah bagaimana mengintegrasikan teknik segmentasi SCW dalam sistem LPR dapat dioperasikan dalam kondisi outdoor jika parameter-parameter uji yang digunakan diatur secara baik, sehingga dapat menjawab permasalahan yang menjadi kendala dalam LPR. Metode yang digunakan dalam tugas akhir ini merupakan integrasi dari berbagai tahapan proses yang terdiri dari pra-processing (grayscale), metode segmentasi (SCW, image masking (Sauvola), dan Connected Component Analysis (aspect ratio, orientasi, dan bilangan Euler). Untuk merealisasikan integrasinya dinyatakan dalam susunan algoritma deteksi plat nomor kendaraan, yang diawali dengan pencarian Region of Interest (RoI).
Setelah dilakukan pengujian pada dataset citra uji yaitu yang terdiri dari 20 citra plat nomor, maka dapat diambil kesimpulan bahwa berhasil terdeteksi sebanyak 14 buah (70%), terdeteksi sebagian sebesar 1 buah (5%), dan tidak terdeteksi sama sekali sebanyak 5 buah (25%).
================================================================================================
Intelligent transportation systems are increasingly becoming a necessity nowadays, so that License Plate-Recognition (LPR) techniques and algorithms various continue to be developed for applications to recognize motorized license plates. However, many obstacles are still encountered due to nonstandard number plate conditions (color, shape, size and pattern), weather factors, lighting levels, and resolution of the scan camera also determine the influence of the quality of the image to be detected.
The focus of this final project is how to integrate the SCW’s segmentation technique in the LPR’s system can be operated in outdoor conditions if the test parameters used are set properly, so that it can answer the problems that are constraints in LPR. The method used in this thesis is an integration of various stages of the process consisting of pre-processing (grayscale), segmentation methods (SCW, image masking (Sauvola), and Connected Component Analysis (aspect ratio, orientation, and Euler numbers)). For realizing its integration is expressed in the arrangement of vehicle number plate detection algorithms, which begins with the search for Region of Interest (RoI).
After testing on the test image dataset, which consists of 20 number plate images, it can be concluded that 14 detected (70%) were detected, semi-detected by 1 pieces (5%), and 5 were not detected at all (25%).

Item Type: Thesis (Undergraduate)
Additional Information: RSE 006.425 Fir d-1 2019
Uncontrolled Keywords: Citra, Deteksi, Plat Nomor Kendaraan, SCW
Subjects: T Technology > TA Engineering (General). Civil engineering (General) > TA1637 Image processing--Digital techniques
T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK5101 Telecommunication
Divisions: Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Electrical Engineering > 20201-(S1) Undergraduate Thesis
Depositing User: Rangga Imantaka Firswandy
Date Deposited: 10 May 2022 03:07
Last Modified: 10 May 2022 03:07
URI: http://repository.its.ac.id/id/eprint/60859

Actions (login required)

View Item View Item