LeafGuard AI: Precision Plant Disease Detection

Agriculture is a cornerstone of global economies, and the health of crops is directly tied to food security and economic stability. Agricultural output is severely threatened by plant diseases, especially in developing nations where access to professional diagnostics is limited. To address this challenge, LeafGuard AI is developed as a deep learning-based system that offers a highly accurate, efficient, and automated method for detecting plant diseases through image recognition. This research explores the individual application of multiple deep learning models, including InceptionV3, MobileNet, a customized Convolutional Neural Network (CNN), and a modified ResNet18 model, each trained separately for classification. A dataset of more than 67,000 RGB images of healthy and diseased leaves, sourced from Kaggle and classified into 29 categories, was used for training. The custom CNN achieved 95% accuracy, ResNet18 reached 96%, and MobileNet obtained 94%, while InceptionV3 performed slightly lower at 87%. This study examines the methodologies, architectures, and comparative performance of these models in plant disease detection. The results demonstrate the effectiveness of deep learning in precision agriculture, promoting sustainable farming practices and enhancing global food security.