| NIM | A1C021074 |
|---|
| Namamhs | SUNU FADLI |
|---|
| Judul Artikel | IDENTIFIKASI PERUBAHAN FISIK BIJI KOPI ROBUSTA TERFERMENTASI YEAST DENGAN METODE PENGOLAHAN CITRA DIGITAL (IMAGE PROCESSING) DAN MACHINE LEARNING |
|---|
| Abstrak (Bhs. Indonesia) | Persaingan ketat di pasar kopi global mendorong produsen untuk menghasilkan produk yang tidak hanya unik, tetapi juga bermutu tinggi dan berkelanjutan. Fermentasi terkontrol menggunakan kultur starter menjadi salah satu pendekatan yang efektif dalam meningkatkan kualitas sensori kopi. Proses fermentasi diketahui dapat memodifikasi sifat fisik dan kimia biji kopi, sehingga dibutuhkan metode analisis yang cepat, efisien, dan aplikatif untuk memantau perubahan mutu. Penelitian ini bertujuan untuk: 1) menganalisis kualitas fisik biji kopi robusta terfermentasi menggunakan image processing, 2) mengembangkan model machine learning untuk mengklasifikasikan biji kopi berdasarkan karakteristik fisik, dan 3) menentukan model terbaik. Penelitian dilakukan di Pusat Riset Teknologi Tepat Guna BRIN pada Oktober 2024–Februari 2025, menggunakan desain dua faktor: waktu fermentasi dan jenis media fermentasi. Sebanyak 13 parameter fisik dianalisis menggunakan image processing. Data diuji menggunakan Kruskal Wallis dan tiga model machine learning: support vector machine (SVM), random forest (RF), dan k-nearest neighbors (KNN). Hasil menunjukkan adanya perubahan fisik signifikan antar perlakuan. Model SVM memberikan akurasi tertinggi, yaitu 80,9% (24 jam), 88,7% (48 jam), dan 78,0% (72 jam), mengungguli RF dan KNN. Temuan ini membuktikan bahwa SVM lebih efektif dalam mengklasifikasikan biji kopi berdasarkan fitur citra digital. |
|---|
| Abtrak (Bhs. Inggris) | The intense competition in the global coffee market drives producers to develop products that are not only unique but also high in quality and aligned with sustainability principles. Controlled fermentation using starter cultures has emerged as an effective approach to enhance the sensory quality of coffee. This fermentation process can alter the physical and chemical properties of coffee beans, thus requiring a rapid, efficient, and applicable analytical method to monitor quality changes. This study aims to: 1) analyze the physical quality of yeast-fermented Robusta coffee beans using image processing techniques, 2) develop machine learning models to classify fermented coffee beans based on their physical characteristics, and 3) determine the best-performing model. The research was conducted at the Center for Appropriate Technology Research, BRIN, from October 2024 to February 2025, using a two-factor experimental design: fermentation time and type of fermentation media. A total of 13 physical parameters were extracted through image processing. Data were analyzed using the Kruskal-Wallis test and three machine learning models: Support Vector Machine (SVM), Random Forest (RF), and K-Nearest Neighbors (KNN). The results showed significant physical differences among treatments. SVM consistently achieved the highest accuracy: 80.9% (24 hours), 88.7% (48 hours), and 78.0% (72 hours), outperforming RF and KNN. These findings indicate that SVM is more effective in classifying coffee beans based on digital image features. |
|---|
| Kata kunci | Robusta, Fermentasi, Kualitas Fisik, Image Processing, Machine Learning, Support Vector Machine, Random Forest, K-Nearest Neighbors |
|---|
| Pembimbing 1 | Susanto Budi Sulistyo |
|---|
| Pembimbing 2 | Seri Intan Kuala |
|---|
| Pembimbing 3 | |
|---|
| Tahun | 2025 |
|---|
| Jumlah Halaman | 43 |
|---|
| Tgl. Entri | 2025-07-28 10:36:46.999774 |
|---|