<![CDATA[Accurate determination of oil palm fresh fruit bunch (FFB) ripeness is crucial to ensure crude palm oil (CPO) quality, yet conventional visual inspection remains subjective and inconsistent. This study proposes a non-destructive ripeness detection system based on a multisensor electronic nose combined with a Random Forest classifier. The system employs five metal oxide semiconductor gas sensors (MQ-2, MQ-3, MQ-4, MQ-5, and MQ-135) integrated with an ESP32 microcontroller to capture volatile organic compounds emitted during fruit ripening. Sensor signals were transformed into seven statistical features, including maximum, minimum, delta, mean, standard deviation, area under the curve, and slope. The dataset was divided into 70% training data and 30% testing data, and model performance was evaluated using a confusion matrix. The results demonstrated an accuracy of 95.3%, precision of 94.8%, recall of 95.1%, and an F1-score of 95.0%. The proposed system successfully classified oil palm fruits into four ripeness levels: unripe, underripe, ripe, and overripe. These findings indicate that the developed electronic nose system provides an objective and reliable approach for oil palm ripeness assessment, with strong potential to support harvesting decisions and quality control in the palm oil industry.]]>
Copyrights © 2026
