Rice Type Classification using Proposed CNN Model

Abstract

Rice is a very important crop that provides nutrition to more than half of the world's population. It is widely grown around the world, and its consumption is widespread in many cultures and cuisines. A comprehensive dataset of 75,000 grain images has been compiled as part of a research initiative. This dataset includes a variety of rice varieties commonly grown in Turkey, including Arborio, Basmati, Ipsala, Jasmine, and Karacadag. The primary goal of this research project is to develop an automated identification system that can differentiate between different rice varieties using a convolutional neural network (CNN) architecture. The model used in this study has many layers and algorithms that allow it to process and analyse large amounts of image data. After five epochs of operation, the CNN architecture achieves an all-time high accuracy rate of 86% after extensive experimentation. The findings of this research project contribute significantly to agricultural advancements and provide a robust and reliable method for accurately classifying rice varieties.