<![CDATA[This study applies the apple leaf disease classification method by combining the ResNet50 architecture as a feature extractor and the Broad Learning System (BLS) as a classifier. The developed system is used to identify five classes of apple leaf diseases, namely Apple scab, Black rot, Cedar apple rust, Healthy and Powdery mildew. The dataset used is 3,298 apple leaf images consisting of 630 Apple Scab images, 621 Black Rot images, 275 Cedar Apple Rust images, 1,645 Healthy images, and 127 Powdery Mildew images. The ResNet50 architecture is able to produce informative image feature representations so that the visual characteristics of each class can be optimally utilized by the BLS algorithm in the classification process. The results of testing the proposed model produce excellent performance of 97%. In addition, the precision, recall and F1-score also each reached 97%, indicating that the model is able to classify accurately and consistently across all classes of apple leaf diseases. Thus, the hybrid method of ResNet50 and BLS has been proven effective and has the potential to be developed as a support system for automatic identification of apple leaf diseases.]]>
Copyrights © 2026
