Plant diseases are significant contributors to crop loss in the global economy, as it has been a significant problem to the food security and livelihoods of farmers. The traditional method is manual diagnosis which has been very slow, subjective and depends on the availability of experts. This paper suggests an integrated AI-based plant disease detection-diagnosis system that is able to combine CNNs with ensemble learning, generative AI, and XAI. Some pretrained CNNs like the VGG16 are used to extract the features of the preprocessed image of the leaves whilst the implementation of an ensemble learning model based on the Random Forest classifier enhances the accuracy of the prediction with improved generalization. The XAI methods such as Grad-cam, LIME and SHAP produce graphical descriptions of the affected areas of leaves and increase transparency and consequently user confidence. Lastly, the generative AI models provide context-related treatment advice based on the disease severity, environmental conditions, and past crop data. The diagnostic accuracy and scalability of smallholder and commercial farming are considerable because of experimental assessment of benchmark datasets. Therefore, the suggested framework can be said to be useful in ensuring sustainable agriculture as well as providing farmers in rural areas with smart and available plant health management tools.
MAMATHA et al. (Thu,) studied this question.