<![CDATA[Black pepper (Piper nigrum L.) is one of Indonesia's leading plantation commodities with significant economic value and an important contribution to the agricultural sector. However, its productivity is frequently reduced due to leaf diseases such as Foot Rot, Pollu Disease, and Slow Decline. Conventional disease identification relies on visual observation by experts, which is time-consuming, subjective, and may delay appropriate disease management. This study aims to develop an image-based classification model for black pepper leaf diseases using the EfficientNetB0 deep learning architecture. The research methodology consists of image dataset collection, image preprocessing, data augmentation, dataset partitioning into training and validation sets, model training, and performance evaluation using accuracy, precision, recall, F1-score, and a confusion matrix. The experimental results demonstrate that the proposed model achieved a training accuracy of 98.17% with a training loss of 0.0856, while the highest validation accuracy reached 88.33% at epoch 18 with a validation loss of 0.3568. These results indicate that EfficientNetB0 is capable of effectively learning the characteristics of black pepper leaf diseases, although slight overfitting was observed. Overall, the proposed model demonstrates strong potential for accurate and efficient disease classification and can serve as a decision-support tool for early disease detection in black pepper cultivation. The implementation of this model is expected to contribute to the advancement of smart agriculture by enabling farmers to identify plant diseases more quickly, accurately, and efficiently.]]>
Copyrights © 2026
