Federated Learning for Collaborative Crop Disease Monitoring in Wheat Production

This study addresses the critical challenge of crop diseases in global agriculture by introducing a novel approach that combines wheat production, crop disease monitoring, and federated learning. Crop diseases pose a severe threat to food security, making early diagnosis and intervention imperative. Wheat, a staple crop with significant health implications, particularly concerning diabetes, requires comprehensive research and monitoring. Federated learning, a privacy-centric machine learning technique, emerges as a promising solution to this problem. It allows collaborative model training across decentralized data sources without exposing raw data, ensuring data privacy and security. The study's results, based on three models, demonstrate its effectiveness: Fast R-CNNs achieved an accuracy of 91%, recall of 93%, and F1-score of 87%; Mask R-CNNs reached an accuracy of 93%, recall of 95%, and F1-score of 88%; RetinaNet achieved an accuracy of 84%, recall of 89%, and F1-score of 74%. These results underscore the potential of this approach in revolutionizing disease detection and classification while contributing to global food security and sustainable agriculture. The future work involves refining the federated learning model, optimizing GATs integration, exploring advanced privacy-preserving strategies, and expanding applications to various crops and disease categories, promising a significant enhancement in the quality of life for individuals impacted by diabetes and further advancements in AIoT-driven agricultural disease monitoring on a global scale.