| dc.description.abstract | Indonesia, with its tropical climate, provides ideal conditions for the growth of various types of plants, including bananas, which are one of the main fruit commodities. Banana production in Indonesia reached 25.96 million tons in 2021, an increase of 5.4% from the previous year. Bananas, with their various ripening stages—unripe, semi-ripe, and ripe—offer different health benefits at each stage. Selecting the appropriate ripeness level is crucial for health and storage purposes. Currently, banana ripeness identification is performed manually through visual observation, which can lead to inconsistencies due to subjectivity. Therefore, this study aims to develop a banana ripeness classification system using a Convolutional Neural Network (CNN) algorithm that analyzes changes in the banana peel color. Based on previous research demonstrating the effectiveness of CNN in object classification, the VGG16 model was chosen for this study. The results indicate that CNN is an effective approach for identifying banana ripeness, with the VGG16 model achieving 100% accuracy, precision, recall, and F1-score. This conclusion affirms that a CNN-based method can provide an objective and consistent way to identify banana ripeness, helping to improve the quality and efficiency in banana processing and storage. The implementation of this system is expected to reduce reliance on manual assessment, enhance the accuracy of ripeness classification, and support the banana industry in Indonesia in producing high-quality products. | en_US |
