Integrated design of an efficient multi spectral imaging and federated learning framework for precision crop disease diagnosis in low-resource farming communities
Scientific Reports, Published online: 10 August 2026; doi:10.1038/s41598-026-53339-0 Integrated design of an efficient multi spectral imaging and federated learning framework for precision crop disease diagnosis in low-resource farming communities
Crop diseases present substantial challenges to productivity in resource-constrained settings, where diagnostic tools and infrastructure are often lacking or insufficient. Traditional crop disease diagnosis methods rely on manual inspection, which is labor-intensive, prone to error, and unable to provide real-time or region-specific insights.
These limitations necessitate the development of advanced diagnostic systems that are scalable, efficient, and capable of delivering actionable insights to empower farmers in mitigating crop losses and enhancing food security.
This research introduces an integrated multi-spectral imaging and machine learning framework designed to transform disease diagnosis and management in low-resource farming communities. At the core of this framework is the 3D Spectral-Spatial Convolutional Neural Network (3D SSCNN), which extracts high-resolution spectral-spatial features from hyperspectral image cubes, achieving an accuracy of around 95% within 0.3 seconds per sample.
The Federated Disease Diagnosis Network (Fed-DiagNet) enables distributed training, enhancing scalability and data privacy to further improve regional model accuracy to approximately 92% and reduce training time by nearly 40%.
The Temporal Progression LSTM component provides dynamic trends in disease progression, achieving 90% accuracy up to a 10-day horizon. By integrating disparate data sources, including hyperspectral imagery, environmental data, and pest observations, the Multimodal Transfer Adaptive Network (MTAN) achieves nearly 93% stress identification accuracy.
A Reinforcement Learning-Based Feedback Optimization (RL-FO) system tailors treatment recommendations to local conditions, optimizing for yield improvement and cost-effectiveness. With this system, diagnostic precision reaches approximately 94%, while maintaining real-time efficiency and scalability to provide actionable insights for effective crop management.
Written by urgent.news from Scientific Reports's reporting — not their text. Machine-written — it may contain errors, so check the original before relying on it.
