Klasifikasi Tingkat Kematangan Pisang Emas Menggunakan Support Vector Machine Berbasis Fitur Warna RGB

Abstract

The ripeness level of Pisang Emas (Cavendish banana) significantly affects its quality and market value, yet manual identification by humans is often subjective and inconsistent. This study proposes an automated banana ripeness classification system using Support Vector Machine (SVM) with color features extracted from the RGB color space. A total of 600 images were collected, consisting of 150 images for each of four classes: unripe, half-ripe, ripe, and overripe. Each image was processed to extract six color features, namely the mean and standard deviation of the Red, Green, and Blue channels. The dataset was split into 80% training data and 20% testing data using a stratified approach. Hyperparameter optimization was performed using Grid Search with 5-fold Cross-Validation to determine the optimal values for the C, gamma, and kernel parameters. The best model, using an RBF kernel, achieved an accuracy of 96.67% on the testing data, correctly classifying the majority of samples across all four ripeness classes. The classification results show that simple RGB-based color features combined with an optimized SVM model are highly effective for distinguishing banana ripeness stages. These findings indicate strong potential for implementing this low-complexity approach in practical agricultural quality control and sorting systems.