Smart Parking System Berbasis Deep Learning & Teknologi RFID

Sihombing, Daniel Rodearman (2024) Smart Parking System Berbasis Deep Learning & Teknologi RFID. Other thesis, Institut Teknologi Sepuluh Nopember.

[thumbnail of 5022201125-Undergraduate_Thesis.pdf] Text
5022201125-Undergraduate_Thesis.pdf - Accepted Version
Restricted to Repository staff only until 1 October 2026.

Download (6MB) | Request a copy

Abstract

Laporan kejahatan pencurian sepeda motor yang terus meningkat setiap hari menjadi kekhawatiran serius. Hal ini dipicu oleh berbagai faktor, termasuk kelalaian dari pemilik sepeda motor, petugas parkir yang kurang bertanggung jawab, dan sistem parkir konvensional yang dinilai tidak cukup aman. Kelemahan utama dari sistem parkir konvensional terletak pada pengawasan dan kontrol akses kendaraan yang tidak memadai. Masalah ini mendorong munculnya berbagai inovasi untuk menciptakan sistem parkir elektronik yang lebih aman. Tugas akhir ini mengusulkan solusi inovatif berupa sistem parkir yang diberi nama "Sistem Pintar Parkir Kendaraan (SPARK)", yang mengintegrasikan teknologi Radio Frequency Identification (RFID) dan sistem pengenalan pelat nomor. Sistem ini menggunakan metode Convolutional Neural Network (CNN) dengan algoritma YOLO untuk deteksi lokasi pelat nomor dan EasyOCR untuk pengenalan karakter pada pelat nomor. Kombinasi kartu RFID dan kamera CCTV dalam prosedur masuk dan keluar parkir menjamin peningkatan keamanan sistem parkir. Dalam penerapan SPARK, diperlukan perangkat seperti RFID Reader RC522, NodeMCU, kamera CCTV, dan laptop untuk mengolah hasil pembacaan citra dari CCTV. SPARK juga dilengkapi dengan dashboard berbasis website yang dapat diakses oleh pengguna (user) dan administrator parkir untuk memudahkan pengelolaan data sirkulasi parkir. Setelah serangkaian pengujian menggunakan berbagai dataset dan skenario, sistem SPARK terbukti memiliki kemampuan deteksi pelat nomor sebesar 100% dengan tingkat pengenalan karakter sebesar 94.75% (similarity). Model terbaik dari hasil pelatihan mencapai precision 97.28% dan recall 99.49%. Pengujian sistem untuk prosedur parkir masuk dan keluar mencatat keberhasilan 100%. Selain itu, pengujian akses dashboard pengguna dengan tiga fitur dan dashboard admin dengan sembilan fitur berfungsi dengan sempurna, mencapai keberhasilan 100%.
=================================================================================================================================
The increasing reports of motorcycle thefts are a growing concern every day. This issue is driven by various factors, including negligence by motorcycle owners, irresponsible parking attendants, and a conventional parking system that is deemed inadequate. The main weakness of the conventional parking system lies in insufficient vehicle surveillance and access control. This problem has spurred various innovations to create safer electronic parking systems. This thesis proposes an innovative solution in the form of a parking system named "Smart Vehicle Parking System (SPARK)", which integrates Radio Frequency Identification (RFID) technology and a license plate recognition system. This system employs the Convolutional Neural Network (CNN) method with the YOLO algorithm for detecting license plate locations and EasyOCR for character recognition on license plates. The combination of RFID cards and CCTV cameras in the entrance and exit procedures ensures enhanced security of the parking system. Implementing SPARK requires devices such as the RFID Reader RC522, NodeMCU, CCTV cameras, and a laptop to process the image readings from the CCTV. SPARK is also equipped with a web-based dashboard accessible by users and parking administrators to facilitate the management of parking circulation data. After a series of tests using various datasets and scenarios, the SPARK system has proven to have a 100% license plate detection capability with a character recognition rate of 94.75% (similarity). The best model from the training results achieved a precision of 97.28% and a recall of 99.49%. Testing of the system for both entry and exit procedures recorded a success rate of 100%. Additionally, tests on user dashboard access with three features and the admin dashboard with nine features functioned perfectly, achieving a 100% success rate.

Item Type: Thesis (Other)
Uncontrolled Keywords: Sistem Parkir Pintar, SPARK, RFID, Convolutional Neural Network, EasyOCR, Pengenalan Pelat Nomor, Smart Parking System, YOLO, License Plate Recognition
Subjects: T Technology > TA Engineering (General). Civil engineering (General) > TA1637 Image processing--Digital techniques
T Technology > TA Engineering (General). Civil engineering (General) > TA1650 Face recognition. Optical pattern recognition.
T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK5105.888 Web sites--Design. Web site development.
T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK7882.P3 Pattern recognition systems
T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK5101 Telecommunication
T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK6570_iPhone (Smartphone). RFID. Mobile communication systems.
Divisions: Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Electrical Engineering > 20201-(S1) Undergraduate Thesis
Depositing User: Daniel Rodearman Sihombing
Date Deposited: 29 Jul 2024 17:03
Last Modified: 29 Jul 2024 17:03
URI: http://repository.its.ac.id/id/eprint/109831

Actions (login required)

View Item View Item