<![CDATA[Cocoa is one of Indonesia’s leading export commodities crucial in supporting the national economy. However, a significant challenge in the post-harvest processing of cocoa lies in the drying stage, where uneven drying often leads to inconsistent bean quality. While previous studies have predominantly focused on classifying cocoa beans based on surface defects such as faded, non-faded, fragmented, moldy, and damaged beans, limited research has addressed classification based on moisture levels—specifically distinguishing between dry and wet beans, which is essential for ensuring optimal fermentation, proper storage, and overall product quality. This study presents a classification model based on a Convolutional Neural Network (CNN) employing the You Only Look Once (YOLO) architecture to detect and classify dry and wet cocoa beans by analyzing visual features, particularly color and shape. A dataset of 2,880 labeled images was compiled and used to train and evaluate the model. The proposed system demonstrated strong performance, achieving an accuracy of 99.8%, a precision of 99.15%, and a recall of 99.8%. These results indicate that the model can be a reliable and efficient tool for detecting moisture content in cocoa beans, thereby enhancing quality control, reducing human subjectivity in the inspection process, and supporting automation in the cocoa processing industry to ensure product consistency and added value in the export market.]]>
Copyrights © 2025
