Smart Agriculture Rice Leaf Disease Detection using Deep Learning Model

Rice, a vital crop in India and diseases affecting its leaves can substantially reduce yield and quality, impacting both the economy and food security. Farmers often struggle to identify diseases accurately, leading to incorrect fertilizer application and soil damage. This study seeks to develop an automated system for detecting rice plant diseases using deep learning techniques, concentrating on the most prevalent conditions: brown spot, leaf blast, and leaf blight. The study utilizes convolutional neural networks (CNNs) with transfer learning to recognize diseases in rice leaf images. Three different models, MobileNetV2, ResNet50, and InceptionV3, were employed and optimized for the classification task. The dataset utilized for training and evaluating the models consisted of 5,932 images obtained from the Kaggle platform. The CNN architecture exhibited exceptional performance, outperforming other models with a peak accuracy of 98.69%.