Stacked Machine Learning Models for Diagnosing Durian Leaf Disease using Hand-crafted Features

Durian is a significant a gricultural product, facing various production challenges, with leaf diseases notably threatening both yield and fruit quality. Recent advancements in artificial intelligence, specifically in ma chine learning, have demonstrated considerable promise for plant disease detection and classification. This study introduces a model f or classifying durian leaf diseases, which uses a hand-crafted feature extraction method to create interpretable features. A stacked architecture combining several machine learning models was utilized to enhance both the accuracy and the robustness of the predictions. The findings reveal that this stacked model substantially surpasses individual baseline models, achieving an impressive accuracy of 90.00%.