Abstract: Rice has become one of the most important crops in Asian countries, so the detection of disease in rice crops is a must to increase production. These diseases are mainly caused by bacteria, viruses, and fungi. We classified mainly three types of diseases such as brown spot, hispa, and leaf blast. In this study, we used Convolutional Neural Network to detect and classify the leaf disease in paddy crop. A CNN is a deep learning algorithm that takes an input image, allots values to various aspects that are present in the image, and differentiates from one another. Here AlexNet and LeNet algorithms are used to train and test the sample images of diseased as well as healthy rice crops. This model shows an accuracy of 97% which is quite high when compared to the existing model. In the existing model, the diseased parts of crops are detected using the pattern recognition method, in which some of the affected portions are left out where the leaf patterns are either unclear or not present. The proposed method examines the image row by column so that no affected spots are left out.

Keywords- RiceLeaf Diseases, AlexNet, LeNet, Convolution Neural Network, Disease detection, Disease classification

| DOI: 10.17148/IJIREEICE.2021.9505