Integrated CNN-SVM Approach for Accurate Radish Leaf Disease Classification: A Comparative Study and Performance Analysis

Radish leaf diseases can have a significant negative influence on agricultural productivity, thus accurate and prompt disease detection is essential for efficient disease management. In this paper, this is an CNN model that can be used to forecast five different radishes leaf illnesses: alternatives Blight, White Rust, the Root Rust of Radish, Radishes Mosaic Virus, plus Radish Phyllody. This model has four convolutional layers, four maximum pooling layers, and one fully connected layer. On a dataset of photos of diseased radish leaf, the model's performance has been evaluated using recall, precision, and F1-score metrics. With precision in the range of 76.19% until 81.82% and remembering extending from 75.00% until 82.05% across individual illness classes, the results showed promising categorization capabilities. The model's overall effectiveness has been shown by its weighted average. The F1 score of 81.45%. The CNN model demonstrates its value as a diagnostic and therapeutic tool for radishes leaf illness with a 92% accuracy rate. These results demonstrate the potential of our methodology to support farmers and specialists in early disease intervention, ultimately improving the health and productivity of radish crops. Its performance and application in the field in agriculture is expected to improve as researchers continue to develop and expand the model.