Advancing Wheat Leaf Disease Classification for Sustainable Agriculture Using EfficientNet

The world-wide threat of leaf disease has a great impact on the food security. The prior identification of these threats affects the quality and the wheat crop productivity. This article proposed an approach that automate the leaf disease classification. It uses hybrid approach of Machine Learning (ML) and Convolutional Neural Network (CNN). The prior identification of illness helps to manage the quality-based food production in significant manner. ML performs several operations such as feature extraction and data augmentation which is a baseline for the performance in early diagnosis. Efficient Net-based CNN architecture is used for classification of leaf images that outperforms in accuracy. This mechanism provides 95.12 accuracy in early diagnosis in test set. It accurately determines the illness of wheat leaf. This work highlights how this proposed approach helps farmers to determine prior diagnosis of illness of wheat leaf. For bigger dataset, it utilises transfer learning and can be used in the applications of agricultural domain.