Diagnosis of Plant Infection for Optimised Authomatic Health Monitoring System Using Vision Transformer Models

Mosaic Virus is highly responsible for the crop loss, especially the pepper, which is considered part of food crop, an essential aspect of world’s agriculture. This work aims at automating the classification of disease severity level-related the Pepper Leaf Mosaic Virus Disease threshold through vision transformer (ViT) model to ensure a critical job. The classification of the severity is more often accompanied by a fairly significant amount of workload and arbitrariness, resulting in equipment innovation in most of the locations. The work applies ViT-based approaches with sufficient precision to appropriately classify a dataset, which contains 10,800 meticulously selected photographs. The present model obtained a commendable average accuracy of 97.25%, which shows that the model was able to detect even smaller differences between the classes of severity. The operational criteria that include the tale of the precision, recall, and F1-score highlight the model’s ability to minimize to Feggsor false Positives with its ability to discriminate cases of viruses at different levels of the spectrum. The more advanced comparison charts made hand methodologies stand at the pinnacle in plant pathology studies. This study is not only beneficial in terms of illuminating the understanding about the Pepper Leaf Mosaic Virus but also gives farmers and researchers a transformative tool that could change the way disease control methods are used. Technologies AI that power ViT model are helping the agriculture industries throughout the world to implement sustainability and resiliency of the systems by providing responses in the right time and place in addressing viral infections.