Optimization of a Hybrid CNN-SVM Model for Pineapple Ripeness Classification

Pineapple ripeness classification plays an important role in maintaining fruit quality and supporting the sorting process. Conventional ripeness assessment is commonly performed manually, which may lead to subjectivity and inconsistent results. This study aims to develop a Hybrid Convolutional Neural Network (CNN) and Support Vector Machine (SVM) model for classifying pineapple ripeness levels based on digital images. The dataset consisted of 3,745 pineapple images categorized into five ripeness classes: unripe, under-ripe, half-ripe, ripe, and rotten. Before model training, the images underwent preprocessing and data augmentation using rotation, horizontal flip, zoom out, brightness adjustment, and darkness adjustment techniques to increase dataset diversity. CNN was employed as a feature extractor to capture the visual characteristics of pineapples, while SVM was used as the classifier to determine the ripeness class. Model performance was evaluated using accuracy, precision, recall, and F1-score metrics. The experimental results showed that the proposed Hybrid CNN-SVM model achieved an accuracy of 92%, outperforming CNN (87%), K-Nearest Neighbor (86%), Random Forest (84%), and SVM (71%). These findings indicate that the combination of CNN and SVM effectively improves pineapple ripeness classification performance. The proposed model has the potential to be implemented as an automated, fast, and objective fruit-sorting support system.