Rice Leaf Disease Identification Using Transfer Learning

Abstract

The ongoing growth of human society necessitates a massive increase in agricultural productivity. Plant diseases are the most significant factor in agriculture that influences the amount and quality of produced crops. There are other plant diseases, however, we'll concentrate on rice plant leaf disease. India produces a large number of rice harvests; however, due to a rice disease, productivity is reduced, resulting in a significant loss for Indian farmers. In general, a farmer may tell if his plant is infected by the disease by looking at it directly. However, this method is not always correct. Plant disease identification may now be done automatically using deep learning thanks to advances in artificial intelligence technology. We used many pre-trained transfer learning models in this study and chose the best of them to develop our deep learning model. By using DenseNet169, we have gained the training accuracy of 98% and validation accuracy of 94%, which shows that DenseNet169 performs much better than others.