A Deep Learning Approach for Multiclass Orange Disease Classification

This research investigates “Machine Learning (ML)” models for the multiclass classification of orange diseases, examining “K-Nearest Neighbors (KNN)”, “Convolutional Neural Network (CNN)” with MobileNetV2 architecture, “Random Forest (RF)” and “Support Vector Machine (SVM)”. The study assesses their performance across Blackspot, Canker, Fresh, and Greening disease classes —using metrics and highlighting the models' strengths. Notably, the CNN model utilizing MobileNetV2 exhibits exceptional accuracy, achieving an impressive 98.22% across all disease categories. The findings emphasize the remarkable accuracy and robustness of CNN in precisely identifying and classifying diverse orange diseases. This suggests its pivotal role in revolutionizing agricultural disease detection systems, underscoring the importance of leveraging advanced deep learning architectures for precise crop health management. The study's outcomes highlight the potential impact of employing sophisticated models in the agriculture sector for effective disease monitoring and crop preservation.