A YOLO and CNN based Approach to Detect and Classify Diseases in Apple Plant Leaves

The agricultural sector plays a pivotal role in sustaining human civilization, with plant diseases having a substantial impact on both crop yield and quality. Early and accurate detection of plant diseases is essential for effective disease management and precision agriculture. In this study, a CNN-based approach for detecting deformities in plant leaves, using deep learning models including YOLO v8 for region of interest (ROI) extraction and VGG16, ResNet50, and InceptionV3, for image classification trained on the Plant Village dataset for the diseases scab, rot, and rust in apples along with healthy leaves has been proposed. The accuracy of the proposed approach was evaluated, and a graphical user interface (GUI) was developed for convenient disease detection. The ResNet50, VGG16, and InceptionV3 models achieved validation accuracy of 64%, 94%, and 96% respectively.