DeepLeaf: Revolutionizing Rice Disease Detection and Classification using Convolutional Neural Networks and Random Forest Hybrid Model
Vinay Kukreja, Rishabh Sharma, Satvik Vats, Manika Manwal
Abstract
Rice hispa disease is a severe risk to agricultural production and can result in considerable reductions in crop yield. It is necessary, to put into practice efficient management techniques, to correctly categorize the disease into its various degrees of intensity. In this study, we present a hybrid model for the multi-classification of rice hispa illness that incorporates convolutional neural networks (CNN) and Random Forest (RF). A dataset consisting of 10,000 photos, each of which represents a distinct degree of intensity, was gathered and pre-processed. The CNN component of the model is responsible for extracting distinguishing characteristics from the pictures, while the RF classifier is in charge of incorporating these characteristics into the final classification. The model displays competitive performance when compared to classic machine learning (ML) techniques and deep learning (DL) models, with an overall accuracy of 97.46%. For each degree of intensity, certain performance measures, such as accuracy, recall, and F1-score, are shown. The confusion matrix analysis provides more evidence that the model can distinguish between different illness states. Our hybrid model provides a potential approach for accurate and reliable multi-classification of the rice hispa disease, which paves the way for enhanced techniques for disease control in rice production.