Rancang Bangun Sistem Pendeteksi Keaslian Uang Kertas Rupiah menggunakan Sinar UV dengan Metode Machine Learning

Zulfiansyah, Alvin Daffa Kumara (2023) Rancang Bangun Sistem Pendeteksi Keaslian Uang Kertas Rupiah menggunakan Sinar UV dengan Metode Machine Learning. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 07111940000079-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 September 2025. Download (3MB) | Request a copy

Abstract

Sistem pendeteksi keaslian uang kertas rupiah merupakan metode yang dibutuhkan karena uang merupakan materi penting sehingga sering terjadi tindak kriminalitas. Metode yang sering digunakan adalah image processing untuk scaling dan color conversion, segmentasi untuk mendapatkan pola tertentu, atau dengan memperhatikan ciri fisik uang kertas rupiah, seperti tekstur dan pola terawang. Namun, metode-metode tersebut dapat dikatakan kurang efektif apabila terjadi peningkatan pemalsuan uang kertas rupiah dalam bentuk fisik. Sehingga penulis merancang sistem counting yang menggunakan sinar ultraviolet untuk menampilkan hidden pattern sebagai fitur khusus mendeteksi keaslian uang dengan metode machine learning dan menambahkan sistem counting buatan sendiri dengan fitur rotation invariant untuk melakukan pendeteksian lebih dari satu lembar uang kertas rupiah. Perpaduan image processing dan machine learning pada penelitian ini mampu memberikan hasil pengujian prototipe dengan metode k-NN dan CNN memberikan persentase keberhasilan prediksi sebesar lebih dari 90% dibandingkan dengan SVM, Random Forest, dan Naive Bayes yang kurang dari 85%. Mekanik prototipe ini mampu melakukan pendeteksian kurang dari 4 detik dengan 10 lembar uang kertas yang dideteksi. Dengan meningkatkan metode detection yang lebih expert, seperti deep learning dengan model yang fleksibel jika ingin digunakan dalam pendeteksi mata uang selain rupiah.
=======================================================================================================================================
The counterfeit detection system for Indonesian banknotes is a necessary method due to the importance of money and the prevalence of criminal activities involving counterfeiting. Commonly used methods include image processing for scaling and color conversion, segmentation to extract specific patterns, and consideration of physical features such as texture and watermark patterns on banknotes. However, these methods can be deemed less effective when there is an increase in physical counterfeit banknotes. Therefore, the author has designed a counting system that utilizes ultraviolet light to reveal hidden patterns as special features for authenticating banknotes using machine learning techniques. Additionally, a custom rotation-invariant counting system has been implemented to detect multiple banknotes. The combination of image processing and machine learning in this research has achieved testing results with more than 90% prediction accuracy using k-NN and CNN methods, compared to SVM, Random Forest, and Naive Bayes which achieved less than 85% accuracy. The prototype's mechanism is capable of detecting 10 banknotes in less than 4 seconds. By enhancing the detection method with more advanced techniques, such as deep learning with flexible models, the system can be applied to detect currencies other than the Indonesian rupiah.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Sinar Ultraviolet, Rupiah, Rotation Invariant, Image Processing, Machine Learning.
Subjects:	Q Science > Q Science (General) > Q325.5 Machine learning. Support vector machines. T Technology > TJ Mechanical engineering and machinery > TJ230 Machine design T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK7878 Electronic instruments
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Electrical Engineering > 20201-(S1) Undergraduate Thesis
Depositing User:	Alvin Daffa Kumara Zulfiansyah
Date Deposited:	24 Jul 2023 07:45
Last Modified:	24 Jul 2023 07:45
URI:	http://repository.its.ac.id/id/eprint/99117

Actions (login required)

View Item