Implementasi Algoritma K-nearest Neighbor dan Principal Component Analysis untuk Klasifikasi Tingkat Kematangan Ceri Kopi Robusta Berdasarkan Warna

Abstract

This study addresses the need for objective and efficient classification of Robusta coffee cherry ripeness, which is crucial for ensuring optimal coffee quality. The research aims to develop an automatic classification system using K-Nearest Neighbor (KNN) and Principal Component Analysis (PCA) based on digital image color features. Employing a quantitative experimental approach, the study utilized 150 digital images of Robusta coffee cherries, categorized into three ripeness levels: unripe, semi-ripe, and ripe. Data were collected using a smartphone camera under controlled lighting, and processed with MATLAB for feature extraction and analysis. The RGB color features were reduced using PCA, and classification was performed with KNN. Evaluation metrics included confusion matrix, accuracy, precision, and recall. The results showed that the system achieved an accuracy of 93.33%, successfully classifying 28 out of 30 test images correctly. These findings indicate that the combination of KNN and PCA provides a reliable and practical solution for automated coffee cherry ripeness classification. The study concludes that this approach can enhance post-harvest processes and support consistent quality in Robusta coffee production.