Federated learning-based UAVs for the diagnosis of Plant Diseases

Abstract

The technological revolution for farmers, especially for the safety of their crops from pests, plays an evident change and convenience for the agriculture industry. The current research presented the classification of different pests using federated learning-based UAVs. The designed scenarios comprise four different sites connected with a global model where different parameters for these sites are received from the local model. State-of-the-art EfficientNet deep model with B03 configurations provides the best accuracy for classifying nine types of pests. The system can achieve an accuracy of 99.55% with the augmentation of images into different angles. The federated learning designed UAVs are the most reliable connection with very less computation power during the classification of pests for the agricultural environment.