Abstract
<jats:p>This chapter investigates the application of deep learning (DL) and explainable artificial intelligence (XAI) for plant disease detection and classification in agriculture. A comprehensive system is developed using the EfficientNetB0 architecture and trained on a dataset of 87,000 leaf images covering 38 disease classes across 14 plant species. The proposed model achieves high performance, with accuracy, precision, and recall scores of 99.69%, 98.27%, and 98.26%, respectively, outperforming established architectures such as MobileNetV2, ResNet-50, and a baseline convolutional neural network. To address the interpretability challenges of deep learning, the system integrates the LIME framework, providing spatially grounded and human-readable explanations for individual predictions. Additionally, the chapter discusses data preprocessing techniques, feature extraction methods, and statistical validation using analysis of variance (ANOVA). Finally, the system is deployed as a mobile application to enable farmers to perform real-time plant disease diagnosis efficiently.</jats:p>