Multiclass Classification of Rice Leaf Disease Using Deep Learning Based Model

Agriculture plays an important role in every individual's life, and crop diseases are a common concern for farmers. Unfortunately, the prevalence of crop diseases is increasing day by day, leading to significant financial losses. Similarly, rice leaf disease is a prevalent issue that impacts the growth and productivity of rice plants. It is triggered by various pathogens, such as bacteria, fungi, and viruses, and can take various forms, including leaf blights, spots, and wilts. To diagnose this disease, it is crucial to inspect the leaves of the rice plant, which may exhibit discolouration, deformities, or lesions. The rice leaf disease can cause a decrease in the quality and quantity of the crop. To mitigate this issue, early detection and accurate identification of diseases and pests in rice plants are necessary. While the traditional method of visually inspecting plants for a disease is still in use, utilizing image processing and machine learning techniques for automatic disease identification in rice plants can be more effective. In this study, a deep learning model, specifically a Convolutional Neural Network (CNN) has been employed to categorize and identify different types of rice leaf diseases. To analyze the outcome, the rice leaf dataset available on Kaggle which includes healthy leaves and various types of leaf diseases has been used. The proposed CNN model has been implemented by setting the optimizer parameter to Adam. Furthermore, with the results, it has been discovered that the proposed technique performs most optimally at a learning rate of 0.0002. To evaluate the model's performance, the accuracy parameter has been used along with the loss metrics, and the results showed that the CNN model, when used with the ADAM optimizer, achieves an accuracy (95.72%) and a loss value (0.1630).