Xing et al. (2026) Integrating UAVs, satellite remote sensing, and machine learning in precision agriculture: pathways to sustainable food production, resource efficiency, and scalable innovation

Identification

• Journal: Frontiers in Agronomy
• Year: 2026
• Date: 2026-01-07
• Authors: Yingyig Xing, Liu Xu-ning, Xiukang Wang
• DOI: 10.3389/fagro.2025.1670380

Research Groups

• Key Laboratory of Applied Ecology of Loess Plateau, College of Life Science, Yan’an University, Yan’an, Shaanxi, China

Short Summary

This review synthesizes the synergistic application of Unmanned Aerial Vehicles (UAVs), satellite remote sensing, and machine learning in precision agriculture to enhance efficiency, optimize resource use, and promote environmental sustainability. It highlights significant improvements in crop yield prediction, resource efficiency (e.g., 20–25% irrigation cost reduction, up to 31 kg ha−1 nitrogen reduction), and disease detection accuracy (>95%), while also identifying persistent challenges like data processing complexities, high computational demands, and the need for scalable, cost-effective solutions.

Objective

• To assess the synergistic impacts of high-resolution UAV data, satellite-scale monitoring, and machine learning analytics on agricultural decision-making processes, providing a foundation for data-driven policy frameworks.
• To quantify empirical gains in resource efficiency (e.g., water, fertilizers), yield prediction accuracy, and environmental outcomes, offering actionable insights for sustainable farm management and regulatory guidance.
• To identify key adoption barriers—including computational costs, data interoperability, and scalability issues—and propose innovation pathways to support policy development and practical implementation, ensuring equitable and widespread adoption across diverse agricultural systems.

Study Configuration

• Spatial Scale: UAVs provide high-resolution (0.1–5 cm per pixel) and centimeter-scale crop surveillance; satellites offer large-scale agricultural data and broad spatial coverage (10–30 meters per pixel) for comprehensive landscape analysis. Integrated systems enable macro-scale field assessments with sub-canopy observational precision.
• Temporal Scale: UAVs enable real-time and frequent temporal assessment; satellite imagery provides data across seasons and regions. AI-driven disease detection can occur 2–3 weeks pre-symptom emergence, with real-time analytics and sub-second latency for in-field decisions.

Methodology and Data

• Models used: Convolutional Neural Networks (CNNs), Random Forests (RFs), Support Vector Machines (SVMs), Recurrent Neural Networks (RNNs), XGBoost, Deep Q-Networks (DQN), Long Short-Term Memory (LSTM), Graph Neural Networks (GNN), Meta-transformer, Multi-CNN-Sequence to Sequence (MCNN-Seq), DSSAT (Decision Support System for Agrotechnology Transfer), mixed-effects models, Q-learning, genetic algorithms, ant colony optimization.
• Data sources: UAV imagery (multispectral, thermal infrared, hyperspectral sensors, e.g., DJI Phantom 4 Multispectral, FLIR Tau2, Headwall Nano-Hyperspec), satellite imagery (optical, multispectral, e.g., Sentinel-2, Landsat; Synthetic Aperture Radar (SAR)), IoT sensor networks (soil moisture, nutrient levels, pH, climate conditions), meteorological data, historical pathogen records, ground truth sensors, randomized controlled trials, benchmark datasets (e.g., WeedMap), raw spectral band data, time-series data.

Main Results

• Resource Efficiency: Irrigation costs reduced by 20–25%; nitrogen application reduced by up to 31 kg ha−1 without compromising productivity; water usage reduced by 18–34% through UAV-mediated precision irrigation; 22% yield improvements with 31% fertilizer input reductions using DQN; 647–1,866 L fuel savings through ant colony-optimized path planning; 73–89% route efficiency improvements with genetic algorithms and RTK-GPS.
• Yield Prediction & Crop Monitoring: Crop yield prediction accuracy improved (e.g., R² = 0.83 for wheat, R² = 0.83 for UAV-satellite data fusion, R² = 0.78 for winter wheat with ensemble ML, R² = 0.875 for potatoes, R² = 0.817 for corn, R² = 0.62 or greater for maize); AI-driven disease detection achieved accuracy exceeding 95% (e.g., Botrytis cinerea in tomatoes, powdery mildew in wheat, downy mildew in grapes) and 81–95% outbreak prediction rates 2–3 weeks pre-symptom emergence.
• Image Analysis: 99.1% segmentation accuracy in aquaponic nutrient deficiency detection (CNNs); improved weed/crop classification AUC scores from [0.576-0.681] to [0.782-0.863] using multispectral UAV data; 13–24% improvement in crop classification accuracy (83% overall, 91% maize-specific) through UAV-satellite data fusion; biomass estimation precision of R²=0.923 (RMSE = 18.8%) and LAI accuracy of R²=0.927 (RMSE = 15.1%).
• Computational Challenges: High computational demands (GPU-intensive ML training 50–200 hours per model, 10–100 terabytes per season data processing); algorithmic generalization issues with 12–18% performance degradation across diverse agroecological zones; UAV operational costs of $500–$2000 per square kilometer for sub-decimeter resolution; CNN models typically require 16–32 GB GPU memory for training.

Contributions

This review synthesizes recent advancements and critical barriers in integrating UAVs, satellite remote sensing, and machine learning for precision agriculture. It provides a comprehensive assessment of the synergistic impacts of these technologies, quantifies empirical gains in resource efficiency and yield prediction accuracy, and identifies key adoption barriers. The paper proposes innovation pathways and policy recommendations, offering a roadmap for future research and policy development aimed at optimizing food production systems.

Funding

• Central Guidance Funds for Local Science and Technology Development Project, Grant No. 2024ZY-JCYJ-02-04.
• Shaanxi Provincial Department of Education Youth innovation team construction research project, Grant No. 22JP101, 21JP141, 23JP189.

Citation

@article{Xing2026Integrating,
  author = {Xing, Yingyig and Xu-ning, Liu and Wang, Xiukang},
  title = {Integrating UAVs, satellite remote sensing, and machine learning in precision agriculture: pathways to sustainable food production, resource efficiency, and scalable innovation},
  journal = {Frontiers in Agronomy},
  year = {2026},
  doi = {10.3389/fagro.2025.1670380},
  url = {https://doi.org/10.3389/fagro.2025.1670380}
}