<![CDATA[Oranges are one of the tropical fruits widely cultivated and consumed in Indonesia. This study aims to develop a digital image classification system to distinguish between Gerga and Kalamansi oranges using the Convolutional Neural Network (CNN) method. The system is designed to assist the automatic identification of fruit types based on visual characteristics found in digital images.The dataset consists of 836 images, including 746 training images, 90 validation images, and 24 testing images. Before the training process, the images undergo preprocessing and data augmentation to increase data variation. The CNN model used in this study consists of several convolutional layers, ReLU activation, pooling, flatten, and fully connected layers for classification.The testing results show an accuracy of 83%, precision of 75%, recall of 100%, and an F1-score of 85%. Overall, the CNN method proves to be sufficiently effective in classifying orange types based on digital images and has the potential to be further developed for automation in the agricultural sector.]]>
Copyrights © 2025
