Lemon Leaf Disease Recognition Using Vision Transformer Networks

Diseases and pests in lemon leaves present significant inadequacies in citrus production, affecting yields and quality. Farmers and experts rely for advice on traditional forms of manual inspection that are slow, laborious and subject to human error. This ongoing innovative research, however, intends to explore the relative effectiveness of state-of-the-art deep learning (DL) models for the identification and classification of lemon leaf conditions automatically. A dataset that includes nine different classes was collected while using various preprocessing techniques for improving the performance of the model. A number of deep learning architectures were considered, such as Vision Transformer (ViT) and Swin Transformer. Impressive results were recorded, whereby ViT achieved 100% accuracy, while Swin Transformer reached 99%. This empirical evidence shows how effective deep learning models are at classifying lemon leaf conditions with perfect accuracy. This, therefore, gives a path to developing reliable and efficient automated systems to aid the farms in disorder and pest management on time, thus increasing citrus production and sustainable agricultural practices.