Paddy leaf disease identification and K-means cluster segmentation using multi-class SVM techniques

The economy of any country is significantly dependent on agricultural yield. A major concern for all countries is facing leaf diseases in agriculture due to bacteria or infections that decrease the yield. To prevent the spreading of leaf diseases, early detection and diagnosis are essential. The plant disease detection technique has been utilized to avoid a reduction in yield percentage. Image processing-based solutions have been quite essential in practice and must be quick, automated, affordable, and precise. The multi-class support vector machine (multi-class SVM) technique has been adopted for this study. This technique extracts information from given samples and provides exceptional results that would help identify and classify the diseases in the plants. Paddy leaf has been adopted for study. This study is focuses clearly on four key components: (i) the research problem early and accurate detection of paddy leaf diseases using image processing; (ii) the methodological approach multi-class SVM classification combined with K-means++ segmentation and complementary descriptors such as color statistics, HOG, and GLCM; (iii) the results achieving an average recognition accuracy of 93 % and peak performance of 96 % with 17 % faster execution compared to traditional methods; and (iv) the practical implications demonstrating the method's potential for efficient, low-cost disease diagnosis in agricultural environments. This strategy has a lot of promise to help with early plant disease diagnosis and enhance crop management techniques. Its implementation has the potential to address a significant gap in agricultural disease management and contribute meaningfully to improving global food security.