Towards Sustainable Mango Cultivation: Automated Severity Classification of Mango Rust Disease using CNN-SVM

The fungal pathogen Puccinia psidii is responsible for the disease known as mango rust, which poses a severe risk to mango farming all over the world. It is essential to efficient illness management to have both accurate and early disease diagnosis as well as a classification of the disease's severity. Convolutional Neural Network (CNN) and Support Vector Machines (SVM) are two types of artificial neural networks. In this study, we provide a unique way to solve this difficulty by introducing a hybrid model that combines the feature extraction skills of CNN with the discriminative capacity of SVM. In the first part of our research, we conduct an exhaustive analysis of conventional approaches to illness evaluation, image-based detection methods, and the revolutionary function that deep learning plays in plant pathology. We provide a description of our methodology, which includes the gathering of data, the preprocessing of that data, and the structure of the hybrid CNN-SVM model. The findings present compelling evidence of the model's superior performance, revealing an overall accuracy of 97.02%. The ability of the model to differentiate between the five unique severity levels of mango rust disease is demonstrated by class-specific metrics such as precision, recall, and F1-score. A comprehensive confusion matrix provides insights into the performance of the model, indicating how well it can reduce the number of incorrect classifications. In addition, our hybrid model outperforms other methods already used in the industry, making it a potentially useful step forward in the mango rust disease severity classification process. The findings of this research have significant consequences, including the elimination of subjectivity and wasteful use of resources in disease management, in addition to contributing to the maintenance of high-quality mango crops.