Klasifikasi Kematangan Buah Apel dengan Ekstraksi Fitur Haralick dan KNN

Abstract

This research aims to classify the ripeness level of apple fruits based on texture features using the Haralick method and color features using histograms. A dataset of 76 apple fruit images was collected. In the preprocessing stage, the apple images were converted to grayscale, followed by the application of a median filter to remove salt and pepper noise, and histogram equalization to enhance image contrast. Texture features were extracted using the Haralick method to obtain contrast, correlation, energy, homogeneity, and entropy features. Color features were extracted using histograms to obtain mean, standard deviation, skewness, and kurtosis. A K-Nearest Neighbor (KNN) model with k = 6 was used for classification. The evaluation results showed an accuracy of 89.47%, precision of 93.75%, recall of 93.75%, and F1-score of 93.75%. This research indicates that texture and color features can effectively classify the ripeness level of apple fruits. Future research can explore more diverse datasets and parameter adjustments to further improve model performance. 

Keywords: apple fruit, ripeness classification, texture features, color features.