Performance Analysis of CNN Models with Data Augmentation in Rice Diseases

Rice is the primary staple crop in the world and an essential food source for millions of people. But a number of diseases that can result in significant quality and productivity losses can affect rice plants. The three most devastating diseases, brown spot, bacterial blight, and leaf blast, have the potential to significantly reduce rice crop production. To solve this issue, a trustworthy and precise automated disease identification system must be developed. In this study, we created a convolution neural network (CNN) model with data augmentation, compared the performance of various CNN models and evaluated its effectiveness in identifying the three elementary diseases that afflict rice plants from leaf images. The accuracy of Xception model, which was 96.67 %, makes it a useful tool for growers to employ in preventing the diseases stated above from affecting their rice the harvesting process.