Harnessing Deep Learning for Tomato Disease Classification

Crop health is one of the most important aspects when it comes to the agriculture industry because tomato crops are very susceptible to several kinds of diseases. Disease diagnosis needs to be thus done in time and with precision for optimal crop output and quality. It implies adopting a hybrid model combining both VGG-19 and ResNet-50 architectures in deep learning framework-based classification of tomato plant diseases. While conventional disease diagnostic methodologies are slow and require an expert, our approach employs residual learning capabilities of ResNet-50 and convolutional layers of VGG-19 in enhancing its robustness as well as accuracy of diagnosis. The first step involves preprocessing, which involves scaling and transformation of grayscale images. The next step involves feature extraction from the image using statistical measurements and texture analysis techniques such as Gray Level Co-Occurrence Matrix (GLCM). In addition, the suggested method uses fine-tuning and transfer learning, utilizing a dataset comprising images that are labeled as healthy and infected tomato leaves. The proposed work has an extended phase of performance evaluation which includes the use of a confusion matrix in order to present the classification outcomes and then applying evaluation metrics such as F1-score, accuracy error rate, precision, and recall. The hybrid model seeks to achieve an estimated accuracy of high precision in classifying photos of tomato leaves into different disease categories. The results demonstrate that deep learning models, particularly the hybrid VGG-19 and ResNet-50 model, can effectively identify subtle disease symptoms, providing a scalable and reliable tool for real-time disease detection in the agricultural sector.