Klasifikasi Mutu dan Kematangan Buah Naga Merah melalui Citra Fisik Buah Menggunakan Metode You Only Look Once Versi 11 Secara Real Time

Abstract

Indonesia is an agricultural country, as its people still rely on the agricultural sector. Red dragon fruit is a superior horticultural commodity in Indonesia with high nutritional content and various health benefits. The super red fleshed variety is most sought after due to its deep color and distinctive flavor. In today's modern era, recognizing and identifying fruit quality and ripeness is very important considering the increasing demand for high-quality fruit. Red dragon fruit, known for its various nutrients and health benefits, is one of the potential commodities in the market. However, many farmers, sellers and buyers still use manual methods to determine quality and ripeness, which results in inefficiency and inaccuracy. Therefore, a system was created using the You Only Look Once (YOLO) method version 11 to classify red dragon fruit based on quality categories (grade A, grade B, and grade C) and ripeness categories (ripe, unripe and rotten) through real-time physical images. In this study, the dataset used was 1800 images, divided into two datasets: 1260 training data, 270 validation data, and 270 testing data. The system built is able to detect and classify fruit in real time with an accuracy of up to 96.30%.