Computer Vision-Based Automated Detection and Severity Grading of Rice Spot Disease for Precision Management

Rice spot disease, a prevalent ailment affecting rice cultivation, presents a substantial challenge to agricultural efficiency. This paper presents a proposed methodology for the automated identification and assessment of rice spot disease using computer vision techniques. The approach uses a combined model consisting of Convolutional Neural Networks (CNN) and Support Vector Machines (SVM) to achieve accurate detection and severity grading. The objective of our study is to facilitate the implementation of precise management strategies for addressing this agricultural concern. By utilizing an extensive dataset comprising images of rice spot disease, we apply cutting-edge methodologies for the preparation of data, augmentation of data, and training of models. The hybrid model employed in our study demonstrates exceptional performance by leveraging the feature extraction capabilities of CNN in conjunction with the classification capability of SVM. The suggested study has achieved an overall accuracy rate of 95.59% in accurately classifying various severity grades of rice spot disease. The observed positive results not only showcase the potential of computer vision and deep learning (DL) in the field of agriculture but also provide a viable and applicable solution for farmers and agricultural professionals. The implementation of computerized disease identification and severity grading enables farmers to make well-informed decisions about crop management, hence resulting in enhanced crop yields and mitigated crop loss. The present study not only makes a valuable contribution to the domain of precision agriculture but also paves the way for future investigations into the use of hybrid models for plant disease identification.