DEEP LEARNING-DRIVEN PLANT DISEASE IDENTIFICATION

Abstract

Plant diseases pose a significant threat to agriculture, making early and accurate detection crucial for maintaining sustainable farming practices. This study introduces an innovative solution by leveraging the advanced capabilities of the Swin Transformer V2 architecture to detect plant diseases. Moving beyond traditional methods, the research explores novel approaches in deep learning, focusing on the Swin Transformer V2's transformative potential. Utilizing the comprehensive Plant Village dataset, the dataset is meticulously fine-tuned and adapted to optimize the performance of the Swin Transformer V2. The model's development is a deliberate and thorough process, ensuring its practical applicability in real-world agricultural scenarios. Through systematic training and rigorous evaluation, the model achieves a remarkable accuracy of 98.2%, surpassing existing models. This significant achievement underscores the model's robustness and effectiveness in identifying plant diseases. The research not only enhances the capabilities of plant disease detection systems but also highlights the transformative impact of advanced deep learning algorithms on sustainable agriculture. By improving disease detection, this study contributes to global food security, showcasing the vital role of cutting-edge technology in addressing agricultural challenges and promoting sustainable farming practices worldwide.