Pengembangan Sistem Prediksi Kualitas Gula Kristal Putih (GKP) Berdasarkan Standar Warna ICUMSA Berbasis Convolutional Neural Network (CNN)

Rafa, Nadya Kautsar (2026) Pengembangan Sistem Prediksi Kualitas Gula Kristal Putih (GKP) Berdasarkan Standar Warna ICUMSA Berbasis Convolutional Neural Network (CNN). Other thesis, Institut Teknologi Sepuluh Nopember.

Text 2042221039-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (3MB) | Request a copy

Abstract

Gula Kristal Putih (GKP) merupakan salah satu komoditas penting di Indonesia yang kualitasnya dinilai berdasarkan parameter warna ICUMSA (International Commission for Uniform Methods of Sugar Analysis). Metode pengukuran ICUMSA secara konvensional bersifat destruktif dan kurang efisien untuk prediksi secara real-time. Penelitian ini bertujuan untuk mengembangkan sistem prediksi berbasis Convolutional Neural Network (CNN) untuk memprediksi nilai ICUMSA pada gula kristal putih melalui analisis citra digital. Sistem dirancang menggunakan Raspberry Pi 5 sebagai kontroler utama, kamera EYD HD Webcam PC03 untuk akuisisi citra, dan LCD Waveshare 7-inch sebagai antarmuka output yang seluruhnya ditempatkan dalam ICUMSA measurement box berukuran 36,5 x 17,5 x 22,5 cm. Citra gula kristal putih diproses dengan model Convolutional Neural Network (CNN). Dataset terdiri dari 80% data training dan 20% data testing, yang kemudian dievaluasi menggunakan metrik Mean Absolute Error (MAE), dan koefisien determinasi (R2). Hasil pengujian menunjukkan bahwa model Convolutional Neural Network (CNN) mampu memprediksi nilai ICUMSA dengan tingkat akurasi yang baik, ditunjukkan oleh nilai MAE sebesar 5,766 IU dan R2 sebesar 0,9104. Selain itu, sistem yang dikembangkan mampu mempercepat proses pengukuran ICUMSA yang secara konvensional memerlukan waktu hingga satu jam atau lebih, menjadi hanya 3-5 detik melalui proses akuisisi citra dan prediksi menggunakan model CNN. Dengan demikian, sistem yang dikembangkan berpotensi menjadi alternatif metode pengukuran ICUMSA yang cepat, non-destruktif, dan praktis untuk mendukung pengendalian mutu gula kristal putih secara real-time di industri.
Kata kunci: ICUMSA, CNN, Raspberry Pi, Gula Kristal Putih, Pengolahan Citra
==================================================================================================================================
White crystal sugar is a vital commodity in Indonesia, with its quality evaluated based on ICUMSA (International Commission for Uniform Methods of Sugar Analysis) color parameters. Conventional ICUMSA measurement methods are destructive, require chemical reagents, and are inefficient for real-time prediction. This research aims to develop a prediction system based on a Convolutional Neural Network (CNN) to predict the ICUMSA value of white crystal sugar through digital image analysis. The system is designed using a Raspberry Pi 5 as the main controller, an EYD HD PC03 Webcam for image acquisition, and a 7-inch Waveshare LCD as the output interface, all housed within a custom ICUMSA measurement box with dimensions of 36.5 cm x 17.5 cm x 22.5 cm. The acquired sugar images are processed using a CNN-based model. The dataset is divided into 80% training data and 20% testing data, and model performance is evaluated using Mean Absolute Error (MAE), and the coefficient of determination (R2). The experimental results show that the CNN model is able to predict ICUMSA values with good accuracy, achieving an MAE of 5.766 IU and R2 of 0.9104. Therefore, the developed system has the potential to serve as an efficient, non-destructive, and practical alternative method for ICUMSA measurement to support real-time quality control in the white crystal sugar industry.
Keywords: ICUMSA, CNN, Raspberry Pi, White Crystal Sugar, Image Processing

Item Type:	Thesis (Other)
Uncontrolled Keywords:	ICUMSA, CNN, Raspberry Pi, Gula Kristal Putih, Pengolahan Citra =========================================================== ICUMSA, CNN, Raspberry Pi, White Crystal Sugar, Image Processing
Subjects:	T Technology > TP Chemical technology > TP382 Sugar--Analysis.
Divisions:	Faculty of Vocational > Instrumentation Engineering
Depositing User:	Nadya Kautsar Rafa
Date Deposited:	05 Aug 2026 02:48
Last Modified:	05 Aug 2026 02:48
URI:	http://repository.its.ac.id/id/eprint/143856

Actions (login required)

View Item