Rice Plant Disease Recognition Based on Convolutional Neural Network with Transfer Learning and Data Augmentation

In recent years, farming has faced various challenges, such as crop failures due to climate, pests or plant diseases, unstable prices of agricultural produce in the market, dependence on specific commodities, and others. The purpose of this research is to develop and test a model for rice plant disease recognition based on image processing and CNN that can identify rice plant diseases with a high level of accuracy. ADAM optimizer, transfer learning, and data augmentation techniques were used to enhance the robustness and performance of the model. In this research paper, transfer learning is utilized by using the Inception V3 and Xception models. Experimental results demonstrated that the application of both transfer learning and data augmentation is a strong technique when used together appropriately. The Xception model performed the best when applying data augmentation, reaching up to 80% accuracy on test from a marginally small amount of rice-leaf dataset.