Classification of Oil Palm Leaf Diseases Using YOLOv8-Nano Algorithm

Abstract

Early detection of diseases in oil palm leaves is crucial to prevent a decline in productivity and to maintain the quality of crop yields. This study aims to develop an automatic classification model for oil palm leaf images using the YOLOv8-Nano algorithm. The dataset used consists of three classes—Healthy, Fungal, and Brown Spot—which were divided into training, validation, and testing sets with a ratio of 80:10:10. The training process was conducted over 10 epochs using image dimensions of 224×224 pixels, leveraging pretrained weights from YOLOv8n-cls. Evaluation results show that the model was able to classify the images perfectly, achieving 100% in accuracy, precision, recall, and F1-score. These findings indicate that YOLOv8-Nano is a lightweight yet highly effective algorithm for the classification task of oil palm leaf images. However, further testing with field data is necessary to ensure the model’s generalization ability in real-world scenarios.