Revolutionizing Rice Farming: Automated Identification and Classification of Rice Leaf Blight Disease Using Deep Learning
The model under consideration exhibits the capacity to function as a mechanism for the timely identification and mitigation of rice leaf blight disease, thereby leading to a substantial enhancement in agricultural productivity and guaranteeing food sustainability.
Abstract
Rice is considered a fundamental dietary component for a significant portion of the global population and is regarded as a crucial crop for ensuring food security on a global scale. Rice crops are susceptible to diverse diseases, among them rice leaf blight disease, which is instigated by the fungus Helminthosporium oryzae. The rice leaf blight disease has the potential to result in substantial reductions in crop yield and poses a significant risk to rice cultivation across various regions globally. The current study introduces a novel approach for the multi-classification of rice leaf blight disease utilizing a hybrid model that combines Convolutional Neural Network (CNN) and Long Short-Term Memory (LSTM) techniques. The model is designed to classify the disease based on six distinct levels of severity. A dataset comprising 6000 images of the rice crop was gathered and subjected to pre-processing and augmentation techniques to enhance the model's efficacy. The proposed model attained a total accuracy of 95.25% in detecting and classifying rice leaf blight disease. Ablation studies were conducted to assess the respective contributions of the CNN and LSTM components toward the model's overall performance. Furthermore, a comparison was made among severity levels to determine the optimal severity level for diagnostic purposes. The model under consideration exhibits the capacity to function as a mechanism for the timely identification and mitigation of rice leaf blight disease, thereby leading to a substantial enhancement in agricultural productivity and guaranteeing food sustainability.