Hybrid EfficientNet-B0 and Vision Transformer Framework for Context-Aware Crop Disease Detection from Agricultural Images

DOI

https://www.doi.org/10.59256/indjcst.20260501034

Published In

Volume 5, Issue 1 (2026)

Published Online

January-April 2026

Pages

243-250

Article Type

Original Article

Check for updates

Crop diseases are one of the major problems in agriculture. It has a significant impact on the productivity of the crops as well as the lives of the farmers. Detection of diseases in plants at an early stage is important in order to avoid damage to the crops and enhance the productivity of agriculture. However, the traditional method of plant disease detection is based on observation, which is a tedious process and might result in incorrect identification of diseases due to a lack of knowledge. This paper suggests an efficient method of plant disease detection using images and the EfficientNet-B0 and ViT .The EfficientNet-B0 model is a Convolutional Neural Network model that is used in the extraction of important features in images related to plant leaves, such as color changes, textures, and diseases, whereas the Vision Transformer model is used in the identification of the relationships between different areas in the images using a self-attention mechanism in order to ensure better accuracy in the classification of the images. The model is trained and validated using different datasets, including PlantVillage, PlantDoc, Vegetable images, etc. Depending upon the disease, fertilizer suggestions will also be provided with images, usage guidelines, dosage, and precautions for proper usage of the fertilizer by the farmers. Moreover, a multilingual voice assistant is also integrated into the system for providing information in English, Hindi, and Telugu, thus making it more user-friendly for farmers models.