Integrating UAVs, satellite remote sensing, and machine learning in precision agriculture: pathways to sustainable food production, resource efficiency, and scalable innovation

Abstract

Precision agriculture has emerged as a pivotal approach to achieving sustainable food production by integrating advanced technologies such as Unmanned Aerial Vehicles (UAVs), satellite remote sensing, and machine learning. This review examines the synergistic application of these technologies in enhancing agricultural efficiency, resource optimization, and environmental sustainability. UAVs enable high-resolution, real-time monitoring of crop health, soil conditions, and pest infestations, while satellite remote sensing provides scalable, large-scale agricultural data for comprehensive landscape analysis. Machine learning algorithms, particularly deep learning models like Convolutional Neural Networks (CNNs) and Random Forests (RFs), process complex datasets to deliver actionable insights for precision decision-making, such as yield prediction, nutrient management, and irrigation optimization. Case studies demonstrate that integrating UAV and satellite data with machine learning improves crop yield prediction accuracy and resource use efficiency, reducing irrigation costs by 20–25% and nitrogen application by up to 31 kg ha⁻¹, without compromising productivity. AI-driven disease detection systems have demonstrated high efficacy, with certain models achieving accuracy exceeding 95% in identifying diseases such as Botrytis cinerea in tomatoes, powdery mildew in wheat, and downy mildew in grapes. However, challenges persist, including data processing complexities, high computational demands, and the need for cost-effective, scalable solutions. The findings underscore the transformative potential of these technologies in advancing sustainable agriculture, while emphasizing the necessity for interdisciplinary collaboration, supportive policies such as subsidies for precision agriculture equipment, streamlined regulations for UAV operations, and open data initiatives for satellite imagery, as well as improved accessibility to key technologies including high-resolution multispectral sensors, cloud computing infrastructure, and scalable machine learning platforms for smallholder farmers. This review provides a roadmap for future research and policy development aimed at optimizing food production systems in the face of climate change and growing population demands.

1 Introduction

Precision agriculture has undergone a profound transformation, shifting from localized mechanization to sophisticated, integrated ecosystem management driven by data. This evolution is fueled by the convergence of Unmanned Aerial Vehicles (UAVs), satellite remote sensing, and machine learning (ML). Early systems relied predominantly on ground-based sensors for discrete parameter monitoring (Barbosa Júnior et al., 2024). In stark contrast, contemporary frameworks leverage synergistic technologies: UAVs equipped with hyperspectral and thermal sensors enable real-time, centimeter-scale crop surveillance (Abbas et al., 2023); satellite platforms (e.g., Sentinel-2, Landsat) facilitate synoptic landscape analysis of critical variables such as soil moisture, phenology, and biotic stressors across vast areas (Messina and Modica, 2022); and advanced ML algorithms, particularly Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs), transform these multi-source data streams into actionable insights for optimizing yield and resource allocation (Araújo et al., 2023).

This integrated technological triad addresses fundamental agricultural challenges. It overcomes spatiotemporal limitations by combining high-resolution UAV monitoring (0.1–5 cm per pixel) for frequent temporal assessment with satellite imagery offering broad spatial coverage (10–30 meters per pixel) across seasons and regions (Awais et al., 2022). Furthermore, deep learning models synthesize complex multidimensional datasets—including spectral, thermal, and topographic information — to quantify crop health, nutrient deficits, and pest risks with high accuracy (>95%) (Qu and Su, 2024). ML-driven analytics systematically extrapolate field-level observations to regional agricultural management strategies, empirically enhancing irrigation efficiency by 20–25% and reducing nitrogen application by 31 kg ha⁻¹ without compromising crop productivity (Ali et al., 2023; da Silva et al., 2020; Wu et al., 2020).

Empirical studies confirm the transformative impact of integrated UAV-satellite-machine learning systems in precision agriculture. UAV-satellite data fusion improves crop yield prediction accuracy to R² = 0.83 (Xu et al., 2020), while AI-driven disease detection achieves 81–95% accuracy in identifying infections 2–3 weeks pre-symptom emergence (González-Rodríguez et al., 2024). AI-driven disease detection systems have demonstrated high efficacy, with certain models achieving accuracy exceeding 95% in identifying diseases such as Botrytis cinerea in tomatoes (González-Rodríguez et al., 2024), powdery mildew in wheat (Wu et al., 2025), and downy mildew in grapes (Nkwocha and Chandel, 2025). These results were validated across diverse ecological regions using multispectral and hyperspectral imagery, with training and testing datasets comprising several thousand to tens of thousands of annotated image samples per disease category. Resource optimization driven by these technologies reduces operational costs by 20–25% while simultaneously advancing environmental sustainability goals (Rennings et al., 2024).

Despite these advancements, persistent challenges hinder widespread adoption. The computational demands are substantial, requiring GPU-intensive ML training (50–200 hours per model) to process large-scale datasets (10–100 terabytes per season). Algorithmic generalization remains problematic, with performance degradation of 12–18% observed particularly in deep learning models such as Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs) when transferred across diverse agroecological zones (Aderele et al., 2025). These models, while highly accurate in localized settings, often struggle with variations in soil type, climate, management practices, and crop phenotypes encountered in new regions. Significant infrastructure gaps also impede adoption, such as high UAV operational costs ($500–$2000 per square kilometer) and sensor interoperability issues that limit accessibility for smallholder farmers (Nhamo et al., 2020).

This review synthesizes recent advances and critical barriers in integrating UAVs, satellite remote sensing, and ML for precision agriculture. It has three primary objectives: (1) to assess the synergistic impacts of high-resolution UAV data, satellite-scale monitoring, and ML analytics on agricultural decision-making processes, providing a foundation for data-driven policy frameworks; (2) to quantify empirical gains in resource efficiency (e.g., water, fertilizers), yield prediction accuracy, and environmental outcomes, thereby offering actionable insights for sustainable farm management and regulatory guidance; and (3) 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.

2 Materials and methods

2.1 Literature search strategy

We conducted a systematic review following the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) framework. The primary databases searched included the Web of Science Core Collection, Scopus, IEEE Xplore, PubMed/MEDLINE, and CAB Abstracts. Our search queries combined Boolean operators with domain-specific keywords to ensure comprehensive coverage of relevant literature. Specifically, we used the following search string: (“precision agriculture*” OR “smart farm*” OR “digital agricult*”) AND (“UAV” OR “drone”) AND (“satellite remote sensing” OR “Sentinel-2”) AND (“AI” OR “machine learning” OR “deep learning”). We applied filters to include only peer-reviewed articles published between 2018 and 2025 and written in English. The initial search returned 2,347 records. After removing duplicates and applying the inclusion and exclusion criteria, 101 studies were deemed eligible for final analysis.

2.2 Inclusion and exclusion criteria

We established specific inclusion and exclusion criteria to select studies relevant to our research objectives. Studies were included if they integrated at least two of the following technologies: UAVs, satellites, and machine learning. Additionally, they needed to report quantifiable metrics such as R², RMSE, or yield improvement and include field-scale validation. Studies were excluded if they presented theoretical models without empirical validation, focused on non-agricultural applications, or used duplicate datasets.

2.3 Data extraction and quality assessment

Extracted parameters from the selected studies were cataloged in a structured matrix. Detail level of UAV or satellite imagery, ranging from 0.1 cm to 30 m per pixel. Algorithm architecture used, such as CNN, SVM, or Random Forest. Predictive accuracy measured by metrics like R², RMSE, or AUC. Reduction in inputs such as water (liters per hectare) or nitrogen (kilograms per hectare). Approaches for field or laboratory verification, including ground truth sensors and randomized controlled trials (RCTs). These gains were observed primarily in studies conducted in temperate cropping systems, such as the North China Plain and the U.S. Corn Belt, under controlled field experiments with integrated UAV-satellite data fusion and machine learning-based decision support systems. We assessed the quality of each study using the JBI Critical Appraisal Tool for observational and experimental studies, evaluating the risk of bias in sampling methods, measurement accuracy, and potential confounding factors.

2.4 Analytical framework

A multi-stage evidence synthesis was employed in our analysis: We tabulated technology configurations, including sensor types and machine learning architectures, to provide an overview of current applications (Supplementary Figure S1). We quantified relationships between technology integration and agricultural outcomes through mixed-effects models. Predictors included spatial resolution and machine learning type, while response variables were yield R² and resource efficiency. We coded emergent challenges such as “data interoperability” and “computational cost” using NVivo 14, ensuring inter-coder reliability with a kappa coefficient greater than 0.85. Quantitative claims, such as a reported 20–25% reduction in irrigation, were cross-validated against primary sources and presented in Table 1.

2.5 Technical validation

To ensure the robustness of our findings, we independently verified key performance metrics from the cited studies: (1) We recalculated NDVI/EVI correlations from raw spectral band data when accessible to confirm the accuracy of vegetation indices. (2) Open-source algorithms, such as Sa et al.’s weed detection CNN, were tested on benchmark datasets like WeedMap to validate their performance. (3) We simulated data processing loads for large-scale applications, specifically for farms spanning 10,000 hectares, using the Google Earth Engine API to assess computational feasibility. These simulations, configured with 8 vCPUs and 32 GB RAM, processed a seasonal composite of Sentinel-2 imagery (approximately 500 GB per 10,000 ha) with a mean execution time of 12.5 ± 2.8 minutes per composite, demonstrating the platform’s capacity for efficient, large-scale agricultural analysis. By integrating these methodological steps, we aimed to provide a comprehensive assessment of the current landscape and potential of integrating UAVs, satellite remote sensing, and machine learning in precision agriculture.

3 Application of different technologies in agriculture

3.1 The application of drones in precision agriculture

The advancement of precision agriculture has firmly established Unmanned Aerial Vehicles (UAVs) as indispensable tools in modern farming practices. UAVs play a critical role in monitoring crop health, evaluating field variables, and implementing targeted crop protection strategies (Karthikeyan et al., 2020). Equipped with multispectral, thermal infrared, and hyperspectral sensors, UAVs enable rapid acquisition of high-resolution imagery, capturing detailed spatial and temporal agricultural data (Raimundo et al., 2021). Sophisticated deep learning algorithms, including Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs), process this data to identify critical indicators such as vegetation cover, water stress, and nutrient deficiencies with high precision (Khan et al., 2022). These machine learning techniques enhance the interpretation of UAV-derived datasets, enabling data-driven decision-making for optimized crop management.

Selecting UAV systems requires careful evaluation of payload capacity, endurance, flight stability, and data processing capabilities to minimize noise and ensure high-quality field data (Almasi et al., 2024). Miniature fixed-wing and multirotor drones are particularly advantageous in precision agriculture due to their maneuverability and ability to hover, facilitating detailed inspections (Nahiyoon et al., 2024). Selecting sensors like hyperspectral cameras, LiDAR, and thermal infrared sensors is equally vital because they improve the accuracy of crop condition assessments when integrated strategically (Karmakar et al., 2024). However, synthesizing data from multiple sensors—such as combining multispectral imagery from a DJI Phantom 4 Multispectral with thermal data from a FLIR Tau2 or hyperspectral data from a Headwall Nano-Hyperspec—poses significant challenges in scalability and analysis, particularly within heterogeneous agricultural environments.

UAVs have revolutionized agricultural monitoring by combining high-resolution imagery with real-time analytics to track crop health, predict yields, and mitigate pest and disease outbreaks (Istiak et al., 2023). Advanced remote sensing techniques, including vegetation index calculations and soil characteristic assessments, leverage multispectral and hyperspectral data to optimize water use efficiency and nutrient management (Mwinuka et al., 2022). Machine learning algorithms further enhance these capabilities by generating detailed farm maps that guide precision field interventions, significantly improving resource efficiency and reducing environmental impacts compared to conventional methods (Goodrich et al., 2023). Despite these advancements, the integration of UAV data with satellite remote sensing for large-scale monitoring continues to pose challenges due to the substantial volume of data and the necessity for highly accurate algorithms (Figure 1). Figure 1 illustrates the synergistic framework for the integration of UAV and satellite remote sensing, emphasizing the complementary data flows—from high-resolution UAV imagery to extensive satellite data—and their convergence within a central analytics system designed for agricultural monitoring and decision-making.

Figure 1

Persistent obstacles include managing vast datasets, ensuring algorithmic reliability, and maintaining data processing efficiency. Inaccuracies in machine learning models could lead to flawed agricultural decisions, underscoring the need for continuous optimization of data integration pipelines (Zualkernan et al., 2023). Exponential growth in UAV-generated data further necessitates advanced processing frameworks to maintain accuracy without compromising scalability. Emerging solutions, such as cloud computing platforms (e.g., Google Earth Engine, AWS) and edge AI systems, are increasingly being adopted to handle these large volumes of data efficiently, enabling real-time analytics and reducing latency for in-field decision support. Additionally, model validation across diverse agro-climatic conditions is essential to ensure robustness (Boursianis et al., 2022).

In summary, UAVs represent a transformative leap in precision agriculture, offering unparalleled capabilities for data collection and analysis. While sensor fusion and machine learning integration have markedly improved agricultural precision, ongoing research must address challenges in data management, algorithmic generalization, and cross-sensor compatibility. Overcoming these barriers will unlock the full potential of UAV technology, driving sustainable and efficient farming practices globally.

3.2 Soil nutrients and crop productivity

Satellite remote sensing has become a cornerstone of modern agricultural monitoring, leveraging high-resolution multispectral imagery and Synthetic Aperture Radar (SAR) to track crop growth, evaluate vegetation health, and predict yields with unprecedented precision (Kordi and Yousefi, 2022). Advanced machine learning algorithms, including Support Vector Machines (SVMs) and Random Forests (RFs), extract critical agronomic parameters from satellite datasets, enabling data-driven decision-making for crop management (Table 1). These models enhance processing efficiency and predictive accuracy, uncovering complex environmental dynamics such as soil moisture variability and pest distribution patterns. Research demonstrates that machine learning-driven data fusion frameworks significantly refine feature selection and interpretation, yielding actionable insights for optimized agricultural practices.

The integration of multi-source satellite data—such as optical, multispectral, and SAR imagery—has revolutionized the observation of agricultural systems. High-resolution optical imagery, for instance, reveals robust correlations between vegetation indices (e.g., EVI2: R² = 0.77; NDVI: R² = 0.71) and banana yields during flowering stages, while crown spread (R² = 0.80) outperforms plant height (R² = 0.16) as a leaf area predictor (Aeberli et al., 2023). Multispectral data supports crop health assessments through reflectance analysis, guiding irrigation and fertilization strategies (de Lima et al., 2021). SAR technology further complements these efforts with all-weather monitoring capabilities, ensuring uninterrupted crop surveillance even under cloud cover (Ahmad et al., 2022). Together, these tools enable precision management across vast agricultural landscapes, from pest control to yield optimization (Wang et al., 2022a).

Despite these advancements, challenges persist in processing large-scale satellite datasets efficiently. The computational demands of integrating heterogeneous data sources and training advanced machine learning models hinder real-time applications and scalability. Addressing these limitations requires prioritizing algorithm enhancements to improve accuracy and generalizability, particularly for diverse agroecological contexts.

The synergy between satellite and UAV data marks a transformative shift in sustainable agriculture. UAVs equipped with hyperspectral sensors (e.g., capturing detailed reflectance spectra across numerous narrow bands) and thermal infrared cameras provide real-time vegetation indices and crop water status data, complementing satellite-derived multispectral (e.g., Sentinel-2’s visible to shortwave infrared bands) and Synthetic Aperture Radar (SAR) insights. Critical preprocessing steps include atmospheric correction, geometric registration to align images from different platforms, radiometric normalization to ensure consistent measurements across sensors, and spatial resampling to harmonize the disparate resolution scales. This refined data fusion enhances pest detection, irrigation scheduling, and nutrient management. However, regional disparities in technological infrastructure—such as limited internet connectivity for data transmission, a lack of high-performance computing resources, and high initial costs for UAV and sensor systems—coupled with skill shortages in data analysis and machine learning model interpretation, pose significant barriers to large-scale implementation. These challenges are particularly acute for smallholder farmers, who face prohibitive equipment costs and limited access to technical training, thereby underscoring the need for cost-effective solutions and targeted capacity-building initiatives.

In summary, satellite remote sensing is indispensable for advancing precision agriculture, offering granular insights into crop dynamics and environmental conditions. While machine learning and data fusion have propelled progress, persistent challenges in computational efficiency, model adaptability, and resource accessibility must be addressed. Future efforts should focus on algorithmic innovation and scalable frameworks to fully harness satellite technology’s potential in fostering sustainable, efficient farming systems.

3.3 Decision analysis supported by machine learning

The integration of machine learning (ML) technologies into precision agriculture has catalyzed a paradigm shift in agricultural data management and decision-making processes. This synthesis evaluates seminal studies in the field, analyzing their methodological innovations, technical contributions, and areas requiring refinement. Through comprehensive examination, we elucidate both the transformative potential and inherent challenges of implementing ML solutions in modern farming systems.

Deep learning architectures demonstrate particular promise, exemplified by Zhao et al. (2020) Multi-CNN-Sequence to Sequence (MCNN-Seq) model for SAR-to-optical image conversion. Achieving R² values of 0.9157-0.9824 across multiple crop types, this approach effectively circumvents cloud obstruction challenges while maintaining operational reliability. Complementary developments include Hamer et al. (2020) papros R package, which integrates spatial regionalization with ML to predict the incidence of specific crop pathogen outbreaks with 68% overall accuracy (and a 91% accuracy in predicting the occurrence of an outbreak event). The model primarily used multispectral satellite imagery, meteorological data, and historical pathogen records as input features for these predictions. These advances underscore deep learning’s capacity to process multidimensional agricultural datasets, enabling real-time decision support systems.

Ensemble learning methodologies show particular efficacy in environmental stress analysis. Du et al. (2024) demonstrate how XGBoost-SHAP integration provides both predictive precision (R²=0.85-0.91) and interpretable insights into ozone pollution impacts on wheat yields. This dual capability addresses a critical industry need for actionable intelligence alongside statistical accuracy.

Field implementations reveal tangible operational benefits. Thompson and Puntel (2020) UAV-based nitrogen management system achieved 18.3 ± 6.1 kg grain/kg N efficiency gains with 31 ± 6.3 kg N ha⁻¹ reductions, demonstrating ML’s capacity for sustainable intensification. González-Rodríguez et al. (2024) report AI-driven disease management systems attaining >95% detection accuracy and 81-95% outbreak prediction rates 2–3 weeks pre-symptom emergence, potentially reducing yield losses by 16% while decreasing pesticide usage.

Despite these advancements, several limitations merit consideration. Performance variability across agricultural contexts (Hamer et al., 2020) suggests need for adaptive model architectures. The data-intensive nature of ML systems—requiring high-resolution UAV, satellite, and IoT sensor inputs (Zhao et al., 2023) —poses significant infrastructure challenges, particularly for resource-constrained operations. Current models also inadequately capture multifactorial stressor interactions, with Du et al. (2024) noting critical gaps in representing combined climate-pollution impacts on crop physiology.

Comparative analysis reveals distinct algorithm tradeoffs: traditional ML approaches (SVMs, RFs) offer computational efficiency and interpretability for structured data processing Zhao et al. (2020), while deep learning excels in unstructured data analysis and spatiotemporal pattern recognition (Figure 2). Future research priorities should emphasize: 1) Developing transfer learning frameworks for cross-regional adaptation, 2) Creating edge computing solutions—such as deploying lightweight models on UAVs or field-based IoT gateways—to reduce cloud dependency, enable real-time in-field decisions with sub-second latency, and minimize data transmission bandwidth, and 3) Establishing multidisciplinary collaborations integrating agronomic expertise with computational innovations. These edge systems, while having less raw processing power than cloud platforms, are critical for time-sensitive applications like real-time pest detection and automated irrigation control, offering a complementary paradigm to cloud-based batch processing.

Figure 2

The machine learning-precision agriculture nexus represents a watershed moment in agricultural technology, enabling unprecedented farm management precision. Realizing this potential at scale requires addressing persistent challenges in model generalization, data infrastructure, and multifactorial system modeling. For instance, a yield prediction model trained on data from the U.S. Corn Belt may experience a performance degradation of 12–18% when directly applied to smallholder systems in East Africa without adaptation (Bohra et al., 2025). Conversely, successful transfer learning applications have been demonstrated by models pre-trained on large, diverse satellite image datasets (e.g., BioMassters), which, when fine-tuned with limited local data, have shown robust performance in estimating above-ground biomass across new, previously unseen regions (Joshi et al., 2024). Strategic investments in adaptive algorithms, distributed computing architectures, and translational research partnerships will be essential to democratize these technologies and build climate-resilient global food systems capable of meeting 21st century demands.

4 Integration of different technologies

4.1 Application of UAV and satellite data fusion

The integration of unmanned aerial vehicle (UAV) technology, satellite remote sensing, and machine learning (ML) architectures has fundamentally redefined precision agriculture capabilities, enabling unprecedented granularity in crop monitoring while driving sustainable production through enhanced decision-making systems. This critical analysis examines key technological implementations, their measurable impacts, and persistent implementation challenges within modern agricultural frameworks.

Recent advancements in multi-source data fusion demonstrate significant operational improvements. Liu et al. (2024) achieved R² values of 0.75 (plant nitrogen accumulation) and 0.60 (nitrogen nutrition index) through canopy dual-ex indicator integration, reducing relative root mean square error (RRMSE) to 28.71% and 24.62% respectively. This hybrid approach capitalizes on UAVs’ high-resolution multispectral imaging (0.1–5 cm/pixel) combined with satellites’ broad-area coverage (10–30 m/pixel), enabling centimeter-scale nutrient management while maintaining hectare-scale operational perspectives. Complementary vegetation index applications (NDVI, GNDVI, SAVI with optimized L parameters) further enhance crop health monitoring precision through multi-spectral pattern recognition (Candiago et al., 2015). However, the technical complexity of heterogeneous data fusion and prohibitive UAV operational costs (~$500–$2000/km² for sub-decimeter resolution) present significant adoption barriers for small-scale agricultural operations. While this cost is substantially higher than traditional ground-based scouting (~$50–$150/km²), the UAV-based approach provides far more frequent, consistent, and high-resolution data, leading to more precise input applications and potential long-term cost savings in water, fertilizers, and pesticides (Guebsi et al., 2024).

Deep learning implementations show particular promise in automated agricultural analytics. Taha et al. (2022) demonstrated 99.1% segmentation accuracy in aquaponic nutrient deficiency detection using convolutional neural networks (CNNs), while Sa et al. (2018) improved weed/crop classification AUC scores from [0.576-0.681] to [0.782-0.863] through multispectral UAV data processing. These architectures enable real-time diagnostics (<2s/image processing) critical for precision nutrient management. Nevertheless, computational requirements remain substantial, with CNN models typically demanding 16–32 GB GPU memory for training – a significant infrastructure hurdle for field deployments.

Integrated ML systems achieve measurable performance gains across agricultural metrics. Hegarty-Craver et al. (2020) reported 13-24% improvement in crop classification accuracy (83% overall, 91% maize-specific) through UAV-satellite data fusion, while Maimaitijiang et al. (2020) attained biomass estimation precision of R²=0.923 (RMSE = 18.8%) and leaf area index (LAI) accuracy of R²=0.927 (RMSE = 15.1%) using hybrid sensor inputs. Multi-Source Remote Sensing Recommendation Algorithm (MRS_AMRA) further enhanced system responsiveness through real-time soil-crop-climate monitoring, achieving 22.5% improvement in NDCG@50 (Normalized Discounted Cumulative Gain at 50, measuring the quality of ranked recommendations) and 12.1% improvement in F1-Score (a balanced measure of classification accuracy combining precision and recall) in adaptive resource allocation (Zhu et al., 2023). These systems nevertheless remain vulnerable to environmental variability – sensor calibration drift (± 5-15% accuracy loss under extreme conditions) and weather-induced data noise require ongoing algorithmic refinement.

Implementation challenges persist across three primary domains: 1) Infrastructure demands, which include a minimum of 25 UAV-sensor setups and seasonal data storage needs of 10–100 TB, as observed in large-scale commercial farming operations (Maimaitijiang et al., 2020); 2) Computational complexity, characterized by model training times ranging from 50 to 200 hours; and 3) Environmental adaptability, which necessitates robust sensor calibration to maintain accuracy across diverse field conditions.

This technological convergence represents a critical pathway toward climate-resilient agriculture, with demonstrated capabilities to enhance nitrogen use efficiency by 18-31% (Thompson and Puntel, 2020) while reducing pesticide applications through 95% accurate disease prediction systems (González-Rodríguez et al., 2024). Combining UAV and satellite data with machine learning has enhanced the accuracy of crop yield predictions and resource utilization efficiency, reducing irrigation costs by 20% to 25% (Zhai et al., 2025), and decreasing nitrogen fertilizer application by up to 31 kg ha⁻¹ without compromising productivity (Scatolini et al., 2024). These results were derived from trials in irrigated maize and wheat systems, utilizing high-resolution multispectral UAV data and Sentinel-2 satellite imagery, under conditions of moderate soil variability and controlled water and nitrogen inputs. Realizing these benefits at scale requires concerted efforts to address current limitations in system affordability, operational complexity, and environmental robustness – challenges that will define the next generation of precision agricultural innovation.

4.2 Innovative applications of machine learning algorithms

The evolution of machine learning (ML) architectures has fundamentally transformed precision agriculture through adaptive learning systems and reinforcement learning paradigms, enabling dynamic optimization of agricultural operations across irrigation, fertilization, and pest management domains (Figure 3). Contemporary implementations demonstrate three critical capabilities: 1) Adaptive neural networks achieving 92-97% sensor data fidelity in continuous crop monitoring, 2) Q-learning agents reducing water usage by 18-34% through UAV-mediated precision irrigation, and 3) Deep reinforcement architectures, specifically Deep Q-Networks (DQN), optimizing fertilizer schedules for maize, resulting in 22% yield improvements at 31% input reductions (Rani et al., 2023). This operational framework establishes a closed-loop system where real-time vegetation indices (NDRE, NDVI) inform reward functions for autonomous field machinery.

Figure 3

Rani et al., 2023). The diagram highlights the integration of machine learning in seed processing, growth monitoring, nutrient management, and yield prediction. Data sources include IoT sensor networks, UAV-based imagery, and satellite-derived environmental variables.

Deep learning breakthroughs are redefining sustainable weed management through multi-spectral image processing. Hu et al. (2024) achieved 99.2% weed discrimination accuracy using hybrid UAV-satellite data fusion, while Li et al. (2021) demonstrated 98.7% variance capture in growth stage clustering through principal component analysis of agricultural management practice (AMP) categories. These architectures leverage 5–10 cm resolution UAV imagery integrated with 10–30 m satellite data, enabling macro-scale field assessments while maintaining sub-canopy observational precision. However, implementation challenges persist, with current models requiring 100–500 training hours on specialized GPU clusters (NVIDIA A100/A6000) and showing 12-18% accuracy degradation when transferred across agroecological zones. To address the generalization gap observed in cross-regional applications—where models such as Convolutional Neural Networks (CNNs), Recurrent Neural Networks (RNNs), and complex ensemble methods exhibit performance declines of 12–18%—several promising strategies have emerged. These strategies include: transfer learning and domain adaptation techniques, which fine-tune pre-trained models using limited local data to better align with regional conditions; multi-region training datasets that incorporate diverse agroecological zones during the initial model development phase; explainable AI (XAI) and feature importance analysis to identify which input variables (e.g., soil properties, climate indices) most influence model performance in new contexts; hybrid modeling approaches that integrate process-based crop models with data-driven machine learning to enhance physiological generalizability; and federated learning frameworks that enable model training across distributed datasets without centralizing sensitive or region-specific data.

Ensemble learning architectures now achieve unprecedented predictive capabilities, exemplified by Li et al. (2022b) winter wheat yield model (R²=0.78) utilizing 17 meteorological and soil parameters. The Deep Crop Mapping (DCM) system represents a quantum leap in crop monitoring, combining LSTM networks with attention mechanisms to attain 85.8% mean kappa scores in corn/soybean mapping – 23-35% superior to traditional USDA-NASS methods (Xu et al., 2020). These advancements are powered by IoT sensor networks generating 10–100 TB/ha seasonal datasets, processed through edge computing architectures with <500ms latency.

Bioinspired algorithms demonstrate particular promise in operational optimization, with Maraveas et al. (2023) reporting 647-1,866 L fuel savings through ant colony-optimized path planning. Contemporary implementations combine genetic algorithms with RTK-GPS systems (2–5 cm positioning accuracy) to achieve 73-89% route efficiency improvements. CNN architectures now process multispectral data (5–12 bands) at 50–100 ha/hour rates, enabling real-time pest detection with 93-97% accuracy (Shah et al., 2023).

Emerging architectures push performance boundaries through novel computational paradigms. Sarkar et al. (2024) meta-transformer/GNN hybrid achieves 97% yield prediction accuracy by modeling 78 environmental covariates, while Chen et al. (2024) demonstrate 19% water use reduction in Xinjiang cotton through DSSAT-integrated reinforcement learning. These systems leverage 10^6–10⁷ parameter models trained on 100–500 node GPU clusters, representing both the potential and computational intensiveness (500-2,000 kWh/training cycle) of next-generation agricultural AI.

The field now faces three critical challenges: 1) Bridging the “generalization gap” (15–30% performance loss in cross-regional transfers), 2) Overcoming computational barriers (50k–200k GPU cluster requirements), and 3) Resolving sensor interoperability issues across 40+ IoT protocols. Future progress hinges on developing energy-efficient neuromorphic chips, creating and adopting standardized agricultural data ontologies and APIs (e.g., OGC SensorThings API, ISO 11783 for agricultural machinery data), and establishing cloud-based MLaaS (Machine Learning as a Service) platforms for global accessibility. Such standards are crucial for enabling seamless data exchange and system integration across diverse agricultural platforms and devices.

4.3 Technology implementation challenges

Recent advancements in UAVs and satellite remote sensing technologies have enabled the unprecedented collection of high-resolution spatial data during agricultural operations. However, limitations in data processing and analytical capabilities have emerged as significant obstacles. The processes of data acquisition, storage, and analysis necessitate high-performance computational resources and sophisticated data processing algorithms. Recent advancements in UAVs and satellite remote sensing technologies have facilitated the unprecedented collection of high-resolution spatial data during agricultural operations (Philip Chen and Zhang, 2014). Integrating Faster R-CNN with federated learning (FL)-based ensemble models, particularly through the utilization of the ResNet50 architecture, which achieves an impressive accuracy of 99%, highlights the potential of advanced machine learning techniques in reducing herbicide usage in agriculture (Ortatas et al., 2024).

The widespread adoption of precision agriculture, enabled by the integration of UAVs, satellite data, and advanced machine learning technologies, is essential for sustainable food production. This is demonstrated by notable improvements in crop yields (15-20%), reductions in costs (25-30%), and enhancements in farming efficiency (20-25%) resulting from the automation of all-terrain vehicles (ATVs) (Padhiary et al., 2024). A survey of 504 Brazilian farmers indicates that while 84% utilize at least one digital technology to enhance productivity, the specialized solutions required for precision agriculture highlight the necessity of advanced technical knowledge and skills among farm managers and technical operators (Bolfe et al., 2020). Research indicates that integrating UAV-mounted multispectral and thermal sensors offers a highly effective solution for the real-time monitoring of crop health and soil quality (Awais et al., 2023). This approach significantly improves water stress management and irrigation efficiency in semi-arid regions. The cross-sectional study, validated by a Cronbach’s Alpha of 0.77 and supported by the results of the KMO and Bartlett’s Test, emphasizes that, despite the significant internal consistency and reliability of the data regarding barriers to the adoption of drones and the Internet of Drones (IoD), there remains a substantial need for rigorous UAV training, certification, and adherence to various regional regulations (Askerbekov et al., 2024). Developing standardized international training curricula—covering safe flight operations, data acquisition protocols, and basic maintenance—alongside competency-based certification frameworks (e.g., modeled after the FAA Part 107 certification or the EU’s UAS Operator Certificate) would significantly enhance operational safety and data quality. Operating UAVs necessitates specialized training, which varies by region. Additionally, machine learning algorithms are inherently complex and require high-performance computing resources. To effectively address these challenges, it is essential to pursue continuous technological optimization and to offer digital skills training for agricultural workers.

As the application of UAVs, satellite data, and machine learning technologies in agriculture evolves, these advanced tools are increasingly integrated into precision agriculture to promote sustainable food production. Machine learning, particularly deep learning, plays a crucial role in predicting and optimizing crop growth environments and assessing crop health due to its ability to manage extensive datasets and identify intricate patterns. In the realm of UAV and satellite remote sensing, high-resolution image analysis facilitates precise monitoring of crop growth and effective field management. However, the implementation of these technologies faces challenges related to regulatory frameworks and ethical considerations. Specifically, policy restrictions may affect UAV flight altitude and frequency, thereby limiting the scope and accuracy of data collection. Additionally, ethical issues surrounding data sharing and privacy protection are paramount, necessitating a careful balance between data security and individual privacy. In developing regions particularly, there is a heightened risk of farmer data misuseenedl as the unauthorized selling of yield and land data to commercial entities, or its use in credit scoring without explicit consentt,led, could potentially exploit smallholder farmers and widen existing socioeconomic disparities. Establishing transparent data governance frameworks that enforce farmer data sovereignty, ensure explicit informed consent, and promote equitable benefit-sharing is therefore critical. Internationally, initiatives such as GAIA-X in Europe aim to create a federated and secure data infrastructure for agriculture, while standards like ISO/IEC 30141 provide a reference architecture for trustworthy IoTxsystems, including agricultural applications. Regionally, the European Unionea Data Act and GDPR set stringent requirements for data access and user rights, whereas countries like China are developing localized data classification and cross-border transfer rules under the Data Security Law. These frameworks emphasize the need for data minimization, purpose limitation, and farmer-centric control, ensuring that digitalization benefits are shared equitably while safeguarding privacy and security. Overall, interdisciplinary collaboration and communication are essential for overcoming these barriers, while ongoing technological innovation and adherence to lawful and compliant policies will collectively enhance the positive impact of precision agriculture in achieving sustainable food production.

5 Mechanism of precision agriculture to improve crop yield

5.1 Measurement and influencing factors of crop yield

Integrating Sentinel-2 remote sensing data with stepwise multilinear models and random forest classification can reliably estimate wheat grain yield prior to harvest, achieving an R² of 0.83 and an overall mapping accuracy of 86% (Segarra et al., 2020). This represents a substantial improvement over traditional Landsat+NDVI approaches, which typically achieve R² values of 0.5–0.65 for similar yield prediction tasks, demonstrating the superior predictive capability of modern ML-driven frameworks. This approach offers detailed insights into regional crop performance, akin to yield monitors attached to harvesters that utilize GPS technology. The integration of remote sensing techniques, utilizing satellite and UAV imagery, offers significant potential to enhance the interplay between energy, water, and agricultural management. This approach facilitates comprehensive monitoring of crop health across extensive areas, thereby fostering interdisciplinary research collaborations and promoting sustainable development (Sanders and Masri, 2016). UAV-based remote sensing technologies, particularly through the application of vegetation indices such as the Soil Adjusted Vegetation Index (SAVI) and the NDVI, are effective in predicting soybean grain yield. This capability provides significant advantages for on-farm management, crop marketing, and policy decision-making (da Silva et al., 2020). Collectively, these methods enhance the precision and comprehensiveness of crop yield measurement.

Crop yield is modulated by a myriad of factors, each exerting a significant influence on the growth and ultimate output of crops (Figure 4). Climatic conditions are paramount, with variables such as temperature, precipitation, and solar radiation directly impacting physiological processes in plants. High-quality soils enhance yield stability and boost average yields by more than 10%, while also reducing climate-related yield declines by up to 21% under RCP 8.5 scenarios projected for 2099 (Qiao et al., 2022). These findings underscore the urgent need to adopt soil improvement strategies in response to climate change. Soil characteristics, including texture, fertility, and moisture content, are also critical as healthy soil supplies essential nutrients and water for optimal plant growth. Precision agriculture technologies employ sensors to monitor soil pH, nutrient levels, and moisture content, allowing farmers to optimize their fertilizer and irrigation management practices (Karthika, 2024). This method not only enhances crop yields but also reduces input costs and minimizes environmental impact, which is essential for promoting sustainable agriculture and ensuring food security.

Figure 4

5.2 Precision farming improves soil quality

The integration of IoT technologies with smart farming techniques in contemporary agriculture significantly enhances resource use efficiency and productivity. This is demonstrated by a simulation accuracy of up to 92% in crop yield predictions, alongside a 25.34% reduction in irrigation costs (Ali et al., 2023). These findings underscore the critical role of IoT and smart farming in optimizing agricultural production and ensuring food security, especially in light of the challenges posed by climate change. The integration of AI and the IoT in modern agriculture, exemplified by the use of yield monitors on harvesters equipped with GPS technology, offers precise real-time data and detailed yield mapping. This integration not only significantly reduces labor costs and environmental impacts but also enhances crop management and sustainability (Sharma and Shivandu, 2024). Remote sensing and ground surveys provide data on crop health and yield, which can be combined to improve the accuracy of yield estimates.

Understanding the impact of environmental stressors on photosynthetic CO₂ assimilation is crucial, as factors such as water scarcity, elevated temperatures, and nutrient deficiencies directly influence plant physiological processes (Morales et al., 2020). These stressors can result in significant variations in crop yield, even when soil quality and crop types are consistent. The integration of biotechnology and nanotechnology in the development of smart fertilizers, along with precision agriculture technologies such as soil pH, nutrient, and moisture sensors, can significantly enhance the efficiency of fertilizers and improve irrigation management (Shanmugavel et al., 2023). Real-time soil moisture data enables precise water application, thereby promoting optimal crop growth. Crop management practices, including fertilization, irrigation, and pest control, significantly influence yield outcomes. Advanced data analysis tools improve these practices by identifying the optimal timing and quantities for interventions. Variable rate technology (VRT) customizes the application of fertilizers and pesticides to address the specific needs of different field zones, thereby fostering healthier crops and enhancing yields. The implementation of VRT is highly adaptable across diverse cropping systems: in perennial fruit trees, it enables precise canopy-specific spraying and nutrient management; for short-duration row crops like vegetables, it facilitates rapid, zone-specific adjustments for nitrogen and herbicides; and in paddy rice systems, it supports differentiated input application that accounts for significant in-field variability in soil properties and water distribution.

Enhancing access to climate information and strengthening governmental support are essential for empowering young farmers in the Hindu Kush Himalayas region to make scientifically informed decisions (Ullah et al., 2024). This empowerment will increase their participation in climate change extension programs and social media engagement, ultimately resulting in significant improvements in farming efficiency and productivity. Machine learning algorithms, particularly RFs, have demonstrated substantial predictive power, achieving R² values of 0.875 for potatoes and 0.817 for corn (Kuradusenge et al., 2023). These algorithms significantly contribute to enhancing food security and climate resilience in low- and middle-income countries, including Rwanda.

Precise irrigation control, enabled by soil moisture sensors, conserves water while effectively meeting crop requirements. Similarly, accurate applications of fertilizers and pesticides reduce excessive usage, thereby lowering costs and minimizing environmental impact. As technology continues to advance, precision agriculture will play an increasingly vital role in addressing challenges related to food security and sustainability. The integration of UAVs, satellite data, and machine learning will further revolutionize farming, ensuring efficient and environmentally friendly practices. This evolution assists farmers in meeting the rising global food demand while conserving natural resources and promoting ecological balance.

5.3 Precision irrigation and fertilization for crop yield

Precision irrigation technologies, which utilize IoT-based remote sensing and telecommunications for real-time monitoring of soil moisture and crop water requirements, significantly improve water efficiency and crop yields (Singh et al., 2023). The precise management of water resources, characterized by optimal deficit irrigation thresholds of 60% field capacity (FC) during the seedling and flowering stages, and 80% FC during the prime fruit stage, minimizes water wastage while ensuring adequate hydration during critical growth phases (Li et al., 2024). This approach significantly enhances both the yield and quality of eggplant, as evidenced by the highest recorded yields of 89.1 t ha⁻¹ and 88.8 t ha⁻¹, along with a maximum water productivity of 19.4 kg m⁻³. The implementation of precision agricultural technologies, including SPAD and UAV-based NDVI, in conjunction with optimal irrigation and nitrogen treatments, significantly enhances maize yields by 37.2%, reaching up to 8.649 Mg ha⁻¹ (Széles et al., 2024). Precision irrigation prevents over-saturation, reducing disease risk and promoting healthy crop growth.

By employing remote sensing technology, farmers can collect real-time data on the spatial and temporal variations in soil conditions, thereby enabling targeted fertilization. This method enhances fertilizer use efficiency by approximately 7%, reduces the likelihood of over-application, and mitigates environmental impacts (Singh et al., 2024). Supplying crops with essential nutrients at the appropriate growth stages directly contributes to yield improvements, with some studies indicating a 15% increase in crop production attributable to precision fertilization (Shah and Wu, 2019). This approach not only enhances economic returns for farmers but also aligns with sustainable agricultural practices by minimizing nutrient runoff and pollution. By optimizing fertilizer usage, precision agriculture fosters higher productivity and environmental stewardship, both of which are essential for sustainably addressing the growing global food demand.

The integration of UAVs and satellite data in precision agriculture, supported by GIS and machine learning, enables real-time, high-resolution monitoring of crop conditions and soil health. Machine learning further augments this capability by analyzing historical and sensor data to forecast crop growth patterns and resource needs. Machine learning enhances the predictive accuracy of crop growth patterns by analyzing historical and sensor data. This is evidenced by an average leaf growth increase of 36.75% in predicted frames, a strong correlation (r = 0.812) with ground truth, and a Structural Similarity Index Measure (SSIM) exceeding 94.60% across all time steps, indicating a very high degree of perceptual similarity and structural fidelity between predicted and actual plant growth imagery (Yasrab et al., 2021). The integration of intelligent decision support systems in Agriculture 4.0 not only optimizes management practices but also significantly enhances crop yield and quality (Zhai et al., 2020). The combined utilization of advanced tools enhances crop growth efficiency and provides farmers with scientific management techniques, thereby facilitating the evolution of modern agriculture. These technological innovations are essential for ensuring food security, particularly in the context of climate change and resource constraints, positioning precision agriculture as a fundamental component of future agricultural development.

6 Sustainability and policy recommendations for precision agriculture

6.1 Integration and innovation of new technology

AI technologies, particularly deep learning algorithms such as CNNs and RNNs, play a crucial role in advancing precision agriculture. These algorithms excel at processing complex agricultural data and identifying patterns related to crop growth and pest infestations. By leveraging big data analytics, AI offers predictive insights and decision-making support throughout all stages of agricultural production. By employing AI to analyze time-series data related to climate conditions and soil properties, the newly developed AI-based drought indices have demonstrated superior performance compared to traditional methods, achieving a correlation coefficient of 0.78 with upper soil moisture (Oyounalsoud et al., 2024). The integration of AI and big data significantly enhances agricultural productivity while minimizing resource wastage. This is evidenced by stakeholders prioritizing commercial value (41%), the importance of data analytics (39%), and the necessity of rapid data processing (36%) to improve overall productivity (Osinga et al., 2022). AI-driven solutions present considerable potential for tackling global challenges, especially climate change, by improving adaptive farming practices. The incorporation of AI and data analytics is crucial for fostering sustainable agricultural innovation, which in turn enhances the efficiency and resilience of food production systems.

By integrating real-time environmental data from sensor networks with UAV-derived vegetation indices and satellite-derived soil characteristics, precise farm management becomes achievable. This integration is evidenced by the enhanced predictability of maize yield, with an R² of 0.62 or greater and a root mean square error of prediction (RMSEP) of 0.35 or less, as demonstrated under controlled experimental conditions (Adewopo et al., 2020). The analysis of multi-source data using advanced machine learning models, such as the CNN-LSTM-Attention model, significantly enhances the accuracy of crop yield predictions. R² values have been recorded at up to 0.80, with error margins notably reduced, as evidenced by RMSE values as low as 188.11 kg ha⁻¹ (Lu et al., 2024).

The swift progress in gene editing technologies is also revolutionizing precision agriculture. Tools like CRISPR-Cas9 allow for precise alterations to crop genomes, enhancing traits such as disease resistance and drought tolerance without affecting other genetic characteristics. This capability facilitates rapid and efficient crop improvement, offering unprecedented convenience and speed. These internal modifications not only enhance plants’ adaptability to environmental changes but also promote personalized, high-efficiency crop production within the context of precision agriculture. However, the application of gene editing in agriculture remains contentious. Ethical and legal debates, along with potential ecological risks and unknown long-term effects, present significant challenges. Divergent regulatory frameworks have emerged globally: countries like the United States regulate specific gene-edited products rather than the process itself under the USDA’s SECURE rule, Argentina has implemented a flexible, product-based approach, while the European Union currently subjects most gene-edited crops to stringent GMO regulations. China is actively developing its own regulatory policies to foster innovation while ensuring biosafety. Therefore, it is essential to establish robust regulatory frameworks, conduct comprehensive risk assessments, and foster public acceptance to enable the widespread adoption of gene editing technologies in agriculture. Addressing these issues is critical for unlocking the full potential of gene editing and ensuring its positive contribution to sustainable agricultural practices and global food security.

6.2 Policy-driven and market demand

In response to global food security challenges and the necessity for sustainable agriculture, precision agriculture technologies are increasingly being adopted, driven by policy regulations and market demands. The integration of drones (UAVs) and satellite data, in conjunction with machine learning, is revolutionizing agricultural practices (Wang et al., 2024a). Government policies promote the advancement of agricultural nanotechnology by facilitating the targeted application of nanofertilizers and nanopesticides (Zain et al., 2024). This strategy not only enhances crop quality and nutrient use efficiency but also reduces environmental impact and minimizes chemical waste. These advancements also support the development of smart machinery and agricultural data analytics. Consumer demand and industry efficiency drive precision agriculture adoption. Powered by machine learning, deep learning, the IoT, and remote sensing technologies, the digital transformation of agriculture significantly enhances efficiency and productivity. It offers data-driven insights, optimizes resource allocation, and contributes to ensuring food security in countries such as Ethiopia (Benti et al., 2024). Consequently, the future of precision agriculture is anticipated to witness ongoing advancements in digital, intelligent, and refined production, driven by both policy and market forces. These advancements are underpinned by significant increases in global R&D investment, which grew by approximately 85% in the agricultural technology sector over the past five years, accelerating innovation in smart machinery and data analytics (Wang et al., 2025).

With the global population continuing to rise, precision agriculture has become essential for ensuring food security and sustainability. The integration of drones and satellite data facilitates real-time, high-resolution monitoring, significantly advancing the field of precision agriculture. Despite initial high engagement, the sustained adoption of precision agriculture (PA) technologies among farmers in the North China Plain remains low, currently at 12.0% (Li et al., 2020). However, the growing market demand for efficient water conservation and reduced chemical usage, combined with government and cooperative support, indicates a promising future for these technologies. Notably, 72.8% of farmers have expressed a willingness to adopt PA within the next five years. The integration of advanced algorithms for forecasting diseases and pest infestations, along with digital decision-making tools, enables timely adjustments in irrigation and fertilization strategies. This approach not only enhances crop protection but also promotes sustainability, particularly in regions such as the Near East and North Africa, where rapid phytosanitary control is crucial (Hasanaliyeva et al., 2022). The utilization of high-performance computing platforms, especially the cloud-based model provided by Google Earth Engine, significantly enhances crop yields and reduces costs by effectively managing resources and monitoring crops through spatial information. This is exemplified by the potential increases in water productivity for various crops in the Lake Urmia Basin, which range from 80% to 150% (Ghorbanpour et al., 2022). Precision agriculture meets consumer demand for sustainable food, driving its development.

The integration of UAVs and satellite data significantly enhances agricultural monitoring and management by enabling efficient real-time assessments of crop health. This integration ensures optimal input and output, as evidenced by improvements in water productivity and crop yield within the smallholder agriculture sectors of developing countries (Nhamo et al., 2020).

Policymakers must formulate strategies to address market demands by ensuring that advanced technologies, such as automated spraying and intelligent irrigation, are both accessible and affordable for farmers. This approach will enhance nitrogen use efficiency and promote sustainable agriculture, as evidenced by notable improvements in crop productivity and soil management achieved through real-time nitrogen assessments and the integration of IoT technologies (Ravikumar et al., 2024). The application of data analytics and the IoT in agricultural management has become increasingly essential for optimizing resource utilization and achieving sustainable production. A robust policy framework, along with effective market mechanisms, can foster innovation in precision agriculture. This alignment ensures that technology addresses practical application needs while facilitating the digital transformation of the agricultural industry.

6.3 Global response to climate change

The intensification of global climate change presents significant challenges to agriculture, primarily due to the increased frequency of extreme weather events such as droughts, floods, and temperature fluctuations, which disrupt crop growth cycles and threaten food security. Precision agriculture, which employs remote sensing data from UAVs and satellites in conjunction with machine learning algorithms, facilitates the monitoring of crop conditions, the prediction of climate impacts, and the guidance of responsive measures. These technologies enhance efficiency and optimize resource allocation, providing valuable data support for agricultural risk management. However, challenges related to data collection and processing, as well as the precision required from these models, impede widespread adoption. Notable successes include Convolutional Neural Networks (CNNs) that maintain over 90% accuracy in identifying heat stress in wheat using thermal and multispectral imagery, Long Short-Term Memory (LSTM) networks that predict drought impacts on soybean yields with R² > 0.80, and ensemble models like Random Forest that have been used to map flood-induced crop damage with 85% classification accuracy. Therefore, it is essential to improve the resolution of remote sensing, enhance data processing capabilities, and ensure the generalizability of machine learning models across diverse environments for successful global implementation.

Climate change has a significant impact on food supply and agricultural production, highlighting the urgent need for the adoption of resilient agricultural practices. Rising temperatures, altered precipitation patterns, and the increasing frequency of extreme weather events emphasize the necessity of integrating UAVs, satellite data, and machine learning into agricultural systems. These technologies enable timely monitoring of crop health, inform precision irrigation, fertilization, and pest control strategies, and reduce environmental pollution while optimizing resource allocation to enhance yields. By merging theoretical frameworks with practical applications, precision agriculture provides scientific strategies for regional climate adaptation, thereby supporting local agricultural production and informing policy-making.

To address the severe impacts of climate change on agriculture, the development of long-term climate-adaptive models is essential for ensuring sustainable food production. Precision agriculture, which leverages UAVs, satellite data, and machine learning, facilitates real-time crop monitoring and provides precise guidance for irrigation and fertilization. The analysis of long-term satellite data supports agricultural planning by evaluating the effects of climate change on ecosystems. Furthermore, machine learning enhances predictions of crop yields, aids in pest identification, and improves our understanding of plant-environment interactions, thereby facilitating the development of resilient crop varieties. The continuous optimization of these technologies, in conjunction with Geographic Information Systems (GIS) and crop models, enables the simulation of crop growth under various climate scenarios, thereby enhancing agricultural resilience and adaptability.

7 Summary and prospect

In the context of a growing global population, escalating climate change, and dwindling resources, precision agriculture (PA) emerges as a vital approach to agricultural management. By leveraging drones, satellite technology, and machine learning, PA aims to enhance production efficiency, minimize resource waste, and promote sustainable growth. The future success of PA will depend on significant technological advancements and strategic implementation. Drones are expected to evolve further, integrating advanced sensors and imaging technologies such as thermal imaging, hyperspectral imaging, and LiDAR to provide detailed field data. These drones will increasingly operate autonomously, utilizing optimized flight paths and enhanced data collection capabilities, thereby improving operational efficiency. Concurrently, the use of satellite data is anticipated to expand, offering higher resolution images and real-time updates, which will enable farmers to make timely decisions regarding irrigation, fertilization, and other critical management tasks.

Machine learning and AI are poised to play a crucial role in advancing precision agriculture by enhancing data analysis and providing tailored farming solutions. As machine learning algorithms continue to evolve, they will increasingly manage and analyze complex datasets, thereby improving the accuracy of predictions related to crop growth trends and disease outbreaks. This enhanced analytical capability will empower agricultural managers to develop customized crop management plans that address the specific needs of diverse fields. Furthermore, AI-driven insights from platforms such as the Climate FieldView digital agriculture platform, the IBM Watson Decision Platform for Agriculture, and other bespoke systems leveraging Google’s TensorFlow will facilitate more precise and efficient interventions, optimizing resource utilization and boosting overall productivity. The integration of these technologies will foster a more responsive and adaptable agricultural system, equipped to tackle the challenges posed by a growing global population and changing environmental conditions.

To fully realize the potential of precision agriculture, strategic initiatives are essential. Increased investment in research and development (R&D) from both government and private sectors is crucial, particularly in the areas of drone technology, satellite advancements, and machine learning applications. Establishing collaborative platforms between industry and academia can foster innovation and facilitate the practical application of these technologies. Encouraging data sharing and collaboration among various agricultural stakeholders will enhance the comprehensiveness and accuracy of data analysis, thereby strengthening precision agriculture practices. Furthermore, developing specialized training programs within agricultural institutions will ensure a continuous supply of skilled professionals proficient in data analysis, drone operation, and machine learning. Improving digital literacy among traditional farmers through training and demonstration projects will also be vital, empowering them to effectively utilize these advanced technologies in their farming operations. Lastly, supportive government policies, including financial subsidies and tax incentives, can reduce the cost barriers to adopting these technologies, promoting their widespread use and driving the sustainable advancement of agriculture.

References

• 1

  AbbasA.ZhangZ.ZhengH.AlamiM. M.AlrefaeiA. F.AbbasQ.et al. (2023). Drones in plant disease assessment, efficient monitoring, and detection: a way forward to smart agriculture. Agronomy13, 1524. doi: 10.3390/agronomy13061524

  • CrossRef
  • Google Scholar
  • View reference in article

• 2

  Abdel-salamM.KumarN.MahajanS. (2024). A proposed framework for crop yield prediction using hybrid feature selection approach and optimized machine learning. Neural Computing Appl.36, 20723–20750. doi: 10.1007/s00521-024-10226-x

  • CrossRef
  • Google Scholar
  • View reference in article

• 3

  AdereleM. O.SrivastavaA. K.Butterbach-BahlK.RahimiJ. (2025). Integrating machine learning with agroecosystem modelling: Current state and future challenges. Eur. J. Agron.168, 127610. doi: 10.1016/j.eja.2025.127610

  • CrossRef
  • Google Scholar
  • View reference in article

• 4

  AdewopoJ.PeterH.MohammedI.KamaraA.CraufurdP.VanlauweB. (2020). Can a combination of UAV-derived vegetation indices with biophysical variables improve yield variability assessment in smallholder farms? Agronomy10, 1934. doi: 10.3390/agronomy10121934

  • CrossRef
  • Google Scholar
  • View reference in article

• 5

  AeberliA.PhinnS.JohansenK.RobsonA.LambD. W. (2023). Characterisation of banana plant growth using high-spatiotemporal-resolution multispectral UAV imagery. Remote Sens.15, 679. doi: 10.3390/rs15030679

  • CrossRef
  • Google Scholar
  • View reference in article

• 6

  AhmadU.NasirahmadiA.HenselO.MarinoS. (2022). Technology and data fusion methods to enhance site-specific crop monitoring. Agronomy12, 555. doi: 10.3390/agronomy12030555

  • CrossRef
  • Google Scholar
  • View reference in article

• 7

  AliA.HussainT.TantashutikunN.HussainN.CocettaG. (2023). Application of smart techniques, internet of things and data mining for resource use efficient and sustainable crop production. Agriculture13, 397. doi: 10.3390/agriculture13020397

  • CrossRef
  • Google Scholar
  • View reference in article

• 8

  AlmasiP.XiaoY.PremadasaR.BoyleJ.JaureguiD.WanZ.et al. (2024). A general method for pre-flight preparation in data collection for unmanned aerial vehicle-based bridge inspection. Drones8, 386. doi: 10.3390/drones8080386

  • CrossRef
  • Google Scholar
  • View reference in article

• 9

  AraújoS. O.PeresR. S.RamalhoJ. C.LidonF.BarataJ. (2023). Machine learning applications in agriculture: current trends, challenges, and future perspectives. Agronomy13, 2976. doi: 10.3390/agronomy13122976

  • CrossRef
  • Google Scholar
  • View reference in article

• 10

  AskerbekovD.Garza-ReyesJ. A.Roy GhatakR.JoshiR.KandasamyJ.Luiz de Mattos NascimentoD. (2024). Embracing drones and the Internet of drones systems in manufacturing – An exploration of obstacles. Technol. Soc.78, 102648. doi: 10.1016/j.techsoc.2024.102648

  • CrossRef
  • Google Scholar
  • View reference in article

• 11

  AwaisM.LiW.CheemaM. J. M.HussainS.ShaheenA.AslamB.et al. (2022). Assessment of optimal flying height and timing using high-resolution unmanned aerial vehicle images in precision agriculture. Int. J. Environ. Sci. Technol.19, 2703–2720. doi: 10.1007/s13762-021-03195-4

  • CrossRef
  • Google Scholar
  • View reference in article

• 12

  AwaisM.LiW.CheemaM. J. M.ZamanQ. U.ShaheenA.AslamB.et al. (2023). UAV-based remote sensing in plant stress imagine using high-resolution thermal sensor for digital agriculture practices: a meta-review. Int. J. Environ. Sci. Technol.20, 1135–1152. doi: 10.1007/s13762-021-03801-5

  • CrossRef
  • Google Scholar
  • View reference in article

• 13

  Barbosa JúniorM. R.MoreiraB.CarreiraV.Brito FilhoA.TrentinC.SouzaF.et al. (2024). Precision agriculture in the United States: A comprehensive meta-review inspiring further research, innovation, and adoption. Comput. Electron. Agric.221, 108993. doi: 10.1016/j.compag.2024.108993

  • CrossRef
  • Google Scholar
  • View reference in article

• 14

  BentiN. E.ChakaM. D.SemieA. G.WarkinehB.SoromessaT. (2024). Transforming agriculture with Machine Learning, Deep Learning, and IoT: perspectives from Ethiopia—challenges and opportunities. Discover Agric.2, 63. doi: 10.1007/s44279-024-00066-7

  • CrossRef
  • Google Scholar
  • View reference in article

• 15

  BohraA.NottmeyerS.RenC.ChenS.MaY. (2025). Advancing corn yield mapping in Kenya through transfer learning. Remote Sens.17, 1717. doi: 10.3390/rs17101717

  • CrossRef
  • Google Scholar
  • View reference in article

• 16

  BolfeÉ.L.JorgeL.SanchesI. D. A.Luchiari JúniorA.da CostaC. C.VictoriaD.et al. (2020). Precision and digital agriculture: adoption of technologies and perception of Brazilian farmers. Agriculture10, 653. doi: 10.3390/agriculture10120653

  • CrossRef
  • Google Scholar
  • View reference in article

• 17

  BoursianisA. D.PapadopoulouM. S.DiamantoulakisP.Liopa-TsakalidiA.BarouchasP.SalahasG.et al. (2022). Internet of Things (IoT) and Agricultural Unmanned Aerial Vehicles (UAVs) in smart farming: A comprehensive review. Internet Things18, 100187. doi: 10.1016/j.iot.2020.100187

  • CrossRef
  • Google Scholar
  • View reference in article

• 18

  CandiagoS.RemondinoF.De GiglioM.DubbiniM.GattelliM. (2015). Evaluating multispectral images and vegetation indices for precision farming applications from UAV images. Remote Sens.7, 4026–4047. doi: 10.3390/rs70404026

  • CrossRef
  • Google Scholar
  • View reference in article

• 19

  ChenY.YuZ.HanZ.SunW.HeL. (2024). A decision-making system for cotton irrigation based on reinforcement learning strategy. Agronomy14, 11. doi: 10.3390/agronomy14010011

  • CrossRef
  • Google Scholar
  • View reference in article

• 20

  da SilvaE. E.Rojo BaioF. H.Ribeiro TeodoroL. P.da Silva JuniorC. A.BorgesR. S.TeodoroP. E. (2020). UAV-multispectral and vegetation indices in soybean grain yield prediction based on in situ observation. Remote Sens. Applications: Soc. Environ.18, 100318. doi: 10.1016/j.rsase.2020.100318

  • CrossRef
  • Google Scholar
  • View reference in article

• 21

  de LimaI. P.JorgeR. G.de LimaJ. L. M. P. (2021). Remote sensing monitoring of rice fields: towards assessing water saving irrigation management practices. Front. Remote Sens.2. doi: 10.3389/frsen.2021.762093

  • CrossRef
  • Google Scholar
  • View reference in article

• 22

  DuC.PeiJ.FengZ. (2024). Unraveling the complex interactions between ozone pollution and agricultural productivity in China’s main winter wheat region using an interpretable machine learning framework. Sci. Total Environ.954, 176293. doi: 10.1016/j.scitotenv.2024.176293

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 23

  García-VeraY. E.Polochè-ArangoA.Mendivelso-FajardoC. A.Gutiérrez-BernalF. J. (2024). Hyperspectral image analysis and machine learning techniques for crop disease detection and identification: A review. Sustainability16, 6064. doi: 10.3390/su16146064

  • CrossRef
  • Google Scholar
  • View reference in article

• 24

  GhorbanpourA. K.KisekkaI.AfsharA.HesselsT.TaraghiM.HessariB.et al. (2022). Crop water productivity mapping and benchmarking using remote sensing and google earth engine cloud computing. Remote Sens.14, 4934. doi: 10.3390/rs14194934

  • CrossRef
  • Google Scholar
  • View reference in article

• 25

  González-RodríguezV. E.Izquierdo-BuenoI.CantoralJ. M.CarbúM.GarridoC. (2024). Artificial intelligence: A promising tool for application in phytopathology. Horticulturae10, 197. doi: 10.3390/horticulturae10030197

  • CrossRef
  • Google Scholar
  • View reference in article

• 26

  GoodrichP.BetancourtO.AriasA. C.ZohdiT. (2023). Placement and drone flight path mapping of agricultural soil sensors using machine learning. Comput. Electron. Agric.205, 107591. doi: 10.1016/j.compag.2022.107591

  • CrossRef
  • Google Scholar
  • View reference in article

• 27

  GuebsiR.MamiS.ChokmaniK. (2024). Drones in precision agriculture: A comprehensive review of applications, technologies, and challenges. Drones8, 686. doi: 10.3390/drones8110686

  • CrossRef
  • Google Scholar
  • View reference in article

• 28

  HamerW. B.BirrT.VerreetJ.-A.DuttmannR.KlinkH. (2020). Spatio-temporal prediction of the epidemic spread of dangerous pathogens using machine learning methods. ISPRS Int. J. Geo-Information9, 44. doi: 10.3390/ijgi9010044

  • CrossRef
  • Google Scholar
  • View reference in article

• 29

  HasanaliyevaG.Si AmmourM.YaseenT.RossiV.CaffiT. (2022). Innovations in disease detection and forecasting: A digital roadmap for sustainable management of fruit and foliar disease. Agronomy12, 1707. doi: 10.3390/agronomy12071707

  • CrossRef
  • Google Scholar
  • View reference in article

• 30

  Hegarty-CraverM.PollyJ.O’NeilM.UjenezaN.RineerJ.BeachR. H.et al. (2020). Remote crop mapping at scale: using satellite imagery and UAV-acquired data as ground truth. Remote Sens.12, 1984. doi: 10.3390/rs12121984

  • CrossRef
  • Google Scholar
  • View reference in article

• 31

  HuK.WangZ.ColemanG.BenderA.YaoT.ZengS.et al. (2024). Deep learning techniques for in-crop weed recognition in large-scale grain production systems: a review. Precis. Agric.25, 1–29. doi: 10.1007/s11119-023-10073-1

  • CrossRef
  • Google Scholar

• 32

  IstiakM. A.SyeedM. M. M.HossainM. S.UddinM. F.HasanM.KhanR. H.et al. (2023). Adoption of Unmanned Aerial Vehicle (UAV) imagery in agricultural management: A systematic literature review. Ecol. Inf.78, 102305. doi: 10.1016/j.ecoinf.2023.102305

  • CrossRef
  • Google Scholar
  • View reference in article

• 33

  JoshiA.PradhanB.ChakrabortyS.VaratharajooR.GiteS.AlamriA. (2024). Deep-transfer-learning strategies for crop yield prediction using climate records and satellite image time-series data. Remote Sens.16, 4804. doi: 10.3390/rs16244804

  • CrossRef
  • Google Scholar
  • View reference in article

• 34

  KarmakarP.TengS. W.MurshedM.PangS.LiY.LinH. (2024). Crop monitoring by multimodal remote sensing: A review. Remote Sens. Applications: Soc. Environ.33, 101093. doi: 10.1016/B978-0-323-91068-2.00017-5

  • CrossRef
  • Google Scholar
  • View reference in article

• 35

  KarthikaR. (2024). “ Chapter 8 - Role of precision agriculture in soil fertility and its application to farmers,” in Remote sensing in precision agriculture. Eds. LamineS.SrivastavaP. K.KayadA.Muñoz-ArriolaF.PandeyP. C. ( Academic Press), 137–147.

  • Google Scholar
  • View reference in article

• 36

  KarthikeyanL.ChawlaI.MishraA. K. (2020). A review of remote sensing applications in agriculture for food security: Crop growth and yield, irrigation, and crop losses. J. Hydrology586, 124905. doi: 10.1016/j.jhydrol.2020.124905

  • CrossRef
  • Google Scholar
  • View reference in article

• 37

  KhanA.VibhuteA. D.MaliS.PatilC. H. (2022). A systematic review on hyperspectral imaging technology with a machine and deep learning methodology for agricultural applications. Ecol. Inf.69, 101678. doi: 10.1016/j.ecoinf.2022.101678

  • CrossRef
  • Google Scholar
  • View reference in article

• 38

  KordiF.YousefiH. (2022). Crop classification based on phenology information by using time series of optical and synthetic-aperture radar images. Remote Sens. Applications: Soc. Environ.27, 100812. doi: 10.1016/j.rsase.2022.100812

  • CrossRef
  • Google Scholar
  • View reference in article

• 39

  KuradusengeM.HitimanaE.HanyurwimfuraD.RukundoP.MtongaK.MukasineA.et al. (2023). Crop yield prediction using machine learning models: case of irish potato and maize. Agriculture13, 225. doi: 10.3390/agriculture13010225

  • CrossRef
  • Google Scholar
  • View reference in article

• 40

  LiK.-Y.BurnsideN. G.de LimaR. S.PeciñaM. V.SeppK.Cabral PinheiroV. H.et al. (2021). An automated machine learning framework in unmanned aircraft systems: new insights into agricultural management practices recognition approaches. Remote Sens.13, 3190. doi: 10.3390/rs13163190

  • CrossRef
  • Google Scholar
  • View reference in article

• 41

  LiZ.ChenZ.ChengQ.DuanF.SuiR.HuangX.et al. (2022b). UAV-based hyperspectral and ensemble machine learning for predicting yield in winter wheat. Agronomy12, 202. doi: 10.3390/agronomy12010202

  • CrossRef
  • Google Scholar
  • View reference in article

• 42

  LiW.ClarkB.TaylorJ. A.KendallH.JonesG.LiZ.et al. (2020). A hybrid modelling approach to understanding adoption of precision agriculture technologies in Chinese cropping systems. Comput. Electron. Agric.172, 105305. doi: 10.1016/j.compag.2020.105305

  • CrossRef
  • Google Scholar
  • View reference in article

• 43

  LiY.MiaoY.ZhangJ.CammaranoD.LiS.LiuX.et al. (2022a). Improving estimation of winter wheat nitrogen status using random forest by integrating multi-source data across different agro-ecological zones. Front. Plant Sci.13. doi: 10.3389/fpls.2022.890892

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 44

  LiX.QiangX.YuZ.LiS.SunZ.HeJ.et al. (2024). Effects of different water stresses under subsurface infiltration irrigation on eggplant growth and water productivity. Scientia Hortic.337, 113548. doi: 10.1016/j.scienta.2024.113548

  • CrossRef
  • Google Scholar
  • View reference in article

• 45

  LiP.ZhangH.DengJ.FuL.ChenH.LiC.et al. (2023). Cover crop by irrigation and fertilization improves soil health and maize yield: Establishing a soil health index. Appl. Soil Ecol.182, 104727. doi: 10.1016/j.apsoil.2022.104727

  • CrossRef
  • Google Scholar
  • View reference in article

• 46

  LiuQ.WangC.JiangJ.WuJ.WangX.CaoQ.et al. (2024). Multi-source data fusion improved the potential of proximal fluorescence sensors in predicting nitrogen nutrition status across winter wheat growth stages. Comput. Electron. Agric.219, 108786. doi: 10.1016/j.compag.2024.108786

  • CrossRef
  • Google Scholar
  • View reference in article

• 47

  LuJ.LiJ.FuH.TangX.LiuZ.ChenH.et al. (2024). Deep learning for multi-source data-driven crop yield prediction in northeast China. Agriculture14, 794. doi: 10.3390/agriculture14060794

  • CrossRef
  • Google Scholar
  • View reference in article

• 48

  MaimaitijiangM.SaganV.SidikeP.DaloyeA. M.ErkbolH.FritschiF. B. (2020). Crop monitoring using satellite/UAV data fusion and machine learning. Remote Sens.12, 1357. doi: 10.3390/rs12091357

  • CrossRef
  • Google Scholar
  • View reference in article

• 49

  MaraveasC.AsterisP. G.ArvanitisK. G.BartzanasT.LoukatosD. (2023). Application of bio and nature-inspired algorithms in agricultural engineering. Arch. Comput. Methods Eng.30, 1979–2012. doi: 10.1007/s11831-022-09857-x

  • CrossRef
  • Google Scholar
  • View reference in article

• 50

  MessinaG.ModicaG. (2022). Twenty years of remote sensing applications targeting landscape analysis and environmental issues in olive growing: A review. Remote Sens.14, 5430. doi: 10.3390/rs14215430

  • CrossRef
  • Google Scholar
  • View reference in article

• 51

  MoralesF.AncínM.FakhetD.González-TorralbaJ.GámezA. L.SeminarioA.et al. (2020). Photosynthetic metabolism under stressful growth conditions as a bases for crop breeding and yield improvement. Plants9, 88. doi: 10.3390/plants9010088

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 52

  MwinukaP. R.MouriceS. K.MbunguW. B.MbilinyiB. P.TumboS. D.SchmitterP. (2022). UAV-based multispectral vegetation indices for assessing the interactive effects of water and nitrogen in irrigated horticultural crops production under tropical sub-humid conditions: A case of African eggplant. Agric. Water Manage.266, 107516. doi: 10.1016/j.agwat.2022.107516

  • CrossRef
  • Google Scholar
  • View reference in article

• 53

  NahiyoonS. A.RenZ.WeiP.LiX.LiX.XuJ.et al. (2024). Recent development trends in plant protection UAVs: A journey from conventional practices to cutting-edge technologies—A comprehensive review. Drones8, 457. doi: 10.3390/drones8090457

  • CrossRef
  • Google Scholar
  • View reference in article

• 54

  NhamoL.MagidiJ.NyamugamaA.ClulowA. D.SibandaM.ChimonyoV. G. P.et al. (2020). Prospects of improving agricultural and water productivity through unmanned aerial vehicles. Agriculture10, 256. doi: 10.3390/agriculture10070256

  • CrossRef
  • Google Scholar
  • View reference in article

• 55

  NkwochaC. L.ChandelA. K. (2025). Towards an end-to-end digital framework for precision crop disease diagnosis and management based on emerging sensing and computing technologies: state over past decade and prospects. Computers. 14, 443. doi: 10.3390/computers14100443

  • CrossRef
  • Google Scholar
  • View reference in article

• 56

  OrtatasF. N.OzkayaU.SahinM. E.UlutasH. (2024). Sugar beet farming goes high-tech: a method for automated weed detection using machine learning and deep learning in precision agriculture. Neural Computing Appl.36, 4603–4622. doi: 10.1007/s00521-023-09320-3

  • CrossRef
  • Google Scholar
  • View reference in article

• 57

  OsingaS. A.PaudelD.MouzakitisS. A.AthanasiadisI. N. (2022). Big data in agriculture: Between opportunity and solution. Agric. Syst.195, 103298. doi: 10.1016/j.agsy.2021.103298

  • CrossRef
  • Google Scholar
  • View reference in article

• 58

  OyounalsoudM. S.YilmazA. G.AbdallahM.AbdeljaberA. (2024). Drought prediction using artificial intelligence models based on climate data and soil moisture. Sci. Rep.14, 19700. doi: 10.1038/s41598-024-70406-6

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 59

  PadhiaryM.SahaD.KumarR.SethiL. N.KumarA. (2024). Enhancing precision agriculture: A comprehensive review of machine learning and AI vision applications in all-terrain vehicle for farm automation. Smart Agric. Technol.8, 100483. doi: 10.1016/j.atech.2024.100483

  • CrossRef
  • Google Scholar
  • View reference in article

• 60

  Philip ChenC. L.ZhangC.-Y. (2014). Data-intensive applications, challenges, techniques and technologies: A survey on Big Data. Inf. Sci.275, 314–347. doi: 10.1016/j.ins.2014.01.015

  • CrossRef
  • Google Scholar
  • View reference in article

• 61

  PriyatikantoR.LuY.DashJ.SheffieldJ. (2023). Improving generalisability and transferability of machine-learning-based maize yield prediction model through domain adaptation. Agric. For. Meteorology341, 109652. doi: 10.1016/j.agrformet.2023.109652

  • CrossRef
  • Google Scholar
  • View reference in article

• 62

  QiaoL.WangX.SmithP.FanJ.LuY.EmmettB.et al. (2022). Soil quality both increases crop production and improves resilience to climate change. Nat. Climate Change12, 574–580. doi: 10.1038/s41558-022-01376-8

  • CrossRef
  • Google Scholar
  • View reference in article

• 63

  QuH.-R.SuW.-H. (2024). Deep learning-based weed–crop recognition for smart agricultural equipment: A review. Agronomy14, 363. doi: 10.3390/agronomy14020363

  • CrossRef
  • Google Scholar
  • View reference in article

• 64

  RaiN.ZhangY.RamB. G.SchumacherL.YellavajjalaR. K.BajwaS.et al. (2023). Applications of deep learning in precision weed management: A review. Comput. Electron. Agric.206, 107698. doi: 10.1016/j.compag.2023.107698

  • CrossRef
  • Google Scholar
  • View reference in article

• 65

  RaimundoJ.Lopez-Cuervo MedinaS.PrietoJ. F.Aguirre de MataJ. (2021). Super resolution infrared thermal imaging using pansharpening algorithms: quantitative assessment and application to UAV thermal imaging. Sensors21, 1265. doi: 10.3390/s21041265

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 66

  RaniS.MishraA. K.KatariaA.MallikS.QinH. (2023). Machine learning-based optimal crop selection system in smart agriculture. Sci. Rep.13, 15997. doi: 10.1038/s41598-023-42356-y

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 67

  RavikumarS.VellingiriG.SellaperumalP.PandianK.SivasankarA.SangchulH. (2024). Real-time nitrogen monitoring and management to augment N use efficiency and ecosystem sustainability–A review. J. Hazardous Materials Adv.16, 100466. doi: 10.1016/j.hazadv.2024.100466

  • CrossRef
  • Google Scholar
  • View reference in article

• 68

  RenningsM.BaadenP.BlockC.JohnM.BröringS. (2024). Assessing emerging sustainability-oriented technologies: the case of precision agriculture. Scientometrics129, 2969–2998. doi: 10.1007/s11192-024-05022-2

  • CrossRef
  • Google Scholar
  • View reference in article

• 69

  SaI.PopovićM.KhannaR.ChenZ.LottesP.LiebischF.et al. (2018). WeedMap: A large-scale semantic weed mapping framework using aerial multispectral imaging and deep neural network for precision farming. Remote Sens.10, 1423. doi: 10.3390/rs10091423

  • CrossRef
  • Google Scholar
  • View reference in article

• 70

  SandersK. T.MasriS. F. (2016). The energy-water agriculture nexus: the past, present and future of holistic resource management via remote sensing technologies. J. Cleaner Production117, 73–88. doi: 10.1016/j.jclepro.2016.01.034

  • CrossRef
  • Google Scholar

• 71

  SarkarS.DeyA.PradhanR.SarkarU. M.ChatterjeeC.MondalA.et al. (2024). Crop yield prediction using multimodal meta-transformer and temporal graph neural networks. IEEE Trans. AgriFood Electron.2, 545–553. doi: 10.1109/TAFE.2024.3438330

  • CrossRef
  • Google Scholar
  • View reference in article

• 72

  ScatoliniP.Vaquero-PiñeiroC.CavazzaF.ZucaroR. (2024). Do irrigation water requirements affect crops’ Economic values? Water16, 77. doi: 10.3390/w16010077

  • CrossRef
  • Google Scholar
  • View reference in article

• 73

  SegarraJ.González-TorralbaJ.AranjueloÍ.ArausJ. L.KefauverS. C. (2020). Estimating wheat grain yield using sentinel-2 imagery and exploring topographic features and rainfall effects on wheat performance in navarre, Spain. Remote Sens.12, 2278. doi: 10.3390/rs12142278

  • CrossRef
  • Google Scholar
  • View reference in article

• 74

  ShahS. A.LakhoG. M.KeerioH. A.SattarM. N.HussainG.MehdiM.et al. (2023). Application of drone surveillance for advance agriculture monitoring by android application using convolution neural network. Agronomy13, 1764. doi: 10.3390/agronomy13071764

  • CrossRef
  • Google Scholar
  • View reference in article

• 75

  ShahF.WuW. (2019). Soil and crop management strategies to ensure higher crop productivity within sustainable environments. Sustainability11, 1485. doi: 10.3390/su11051485

  • CrossRef
  • Google Scholar
  • View reference in article

• 76

  ShanmugavelD.RusynI.Solorza-FeriaO.KamarajS.-K. (2023). Sustainable SMART fertilizers in agriculture systems: A review on fundamentals to in-field applications. Sci. Total Environ.904, 166729. doi: 10.1016/j.scitotenv.2023.166729

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 77

  SharmaK.ShivanduS. K. (2024). Integrating artificial intelligence and Internet of Things (IoT) for enhanced crop monitoring and management in precision agriculture. Sensors Int.5, 100292. doi: 10.1016/j.sintl.2024.100292

  • CrossRef
  • Google Scholar
  • View reference in article

• 78

  SinghD.BiswalA. K.SamantaD.SinghV.KadryS.KhanA.et al. (2023). Smart high-yield tomato cultivation: precision irrigation system using the Internet of Things. Front. Plant Sci.14. doi: 10.3389/fpls.2023.1239594

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 79

  SinghS.SinghR.SinghK.KatochK.ZaeenA. A.BirhanD. A.et al. (2024). Smart fertilizer technologies: An environmental impact assessment for sustainable agriculture. Smart Agric. Technol.8, 100504. doi: 10.1016/j.atech.2024.100504

  • CrossRef
  • Google Scholar
  • View reference in article

• 80

  SzélesA.HuzsvaiL.MohammedS.NyékiA.ZagyiP.HorváthÉ.et al. (2024). Precision agricultural technology for advanced monitoring of maize yield under different fertilization and irrigation regimes: A case study in Eastern Hungary (Debrecen). J. Agric. Food Res.15, 100967. doi: 10.1016/j.jafr.2024.100967

  • CrossRef
  • Google Scholar
  • View reference in article

• 81

  TahaM. F.AbdallaA.ElMasryG.GoudaM.ZhouL.ZhaoN.et al. (2022). Using deep convolutional neural network for image-based diagnosis of nutrient deficiencies in plants grown in aquaponics. Chemosensors10, 45. doi: 10.3390/chemosensors10020045

  • CrossRef
  • Google Scholar
  • View reference in article

• 82

  ThompsonL. J.PuntelL. A. (2020). Transforming unmanned aerial vehicle (UAV) and multispectral sensor into a practical decision support system for precision nitrogen management in corn. Remote Sens.12, 1597. doi: 10.3390/rs12101597

  • CrossRef
  • Google Scholar
  • View reference in article

• 83

  UllahA.AdamsF.BavorovaM. (2024). Empowering young farmers’ voices in climate change extension programs: An in-depth analysis of decision-making dynamics and social media engagement. Int. J. Disaster Risk Reduction111, 104713. doi: 10.1016/j.ijdrr.2024.104713

  • CrossRef
  • Google Scholar
  • View reference in article

• 84

  UmutoniL.SamadiV. (2024). Application of machine learning approaches in supporting irrigation decision making: A review. Agric. Water Manage.294, 108710. doi: 10.1016/j.agwat.2024.108710

  • CrossRef
  • Google Scholar
  • View reference in article

• 85

  WangY.AliA.ChenZ. (2025). Dynamic relationships between environment-related technologies, agricultural value added, transport infrastructure and environmental emissions in the five most populous countries. Sci. Rep.15, 2308. doi: 10.1038/s41598-025-86451-8

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 86

  WangD.CaoW.ZhangF.LiZ.XuS.WuX. (2022a). A review of deep learning in multiscale agricultural sensing. Remote Sens.14, 559. doi: 10.3390/rs14030559

  • CrossRef
  • Google Scholar
  • View reference in article

• 87

  WangY.FanL.TaoR.ZhangL.ZhaoW. (2022b). Research on cropping intensity mapping of the Huai River Basin (China) based on multi-source remote sensing data fusion. Environ. Sci. pollut. Res.29, 12661–12679. doi: 10.1007/s11356-021-15387-z

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 88

  WangJ.WangY.LiG.QiZ. (2024a). Integration of remote sensing and machine learning for precision agriculture: A comprehensive perspective on applications. Agronomy14, 1975. doi: 10.3390/agronomy14091975

  • CrossRef
  • Google Scholar
  • View reference in article

• 89

  WangJ.WangY.QiZ. (2024b). Remote sensing data assimilation in crop growth modeling from an agricultural perspective: new insights on challenges and prospects. Agronomy14, 1920. doi: 10.3390/agronomy14091920

  • CrossRef
  • Google Scholar
  • View reference in article

• 90

  WuY.ChenL.YangN.ZongbaoS. (2025). Research progress of deep learning-based artificial intelligence technology in pest and disease detection and control. Agriculture15, 2077. doi: 10.3390/agriculture15192077

  • CrossRef
  • Google Scholar
  • View reference in article

• 91

  WuY.MuG.QinC.MiaoQ.MaW.ZhangX. (2020). Semi-supervised hyperspectral image classification via spatial-regulated self-training. Remote Sens.12, 159. doi: 10.3390/rs12010159

  • CrossRef
  • Google Scholar
  • View reference in article

• 92

  XuJ.ZhuY.ZhongR.LinZ.XuJ.JiangH.et al. (2020). DeepCropMapping: A multi-temporal deep learning approach with improved spatial generalizability for dynamic corn and soybean mapping. Remote Sens. Environ.247, 111946. doi: 10.1016/j.rse.2020.111946

  • CrossRef
  • Google Scholar
  • View reference in article

• 93

  YasrabR.ZhangJ.SmythP.PoundM. P. (2021). Predicting plant growth from time-series data using deep learning. Remote Sens.13, 331. doi: 10.3390/rs13030331

  • CrossRef
  • Google Scholar
  • View reference in article

• 94

  ZainM.MaH.Ur RahmanS.NuruzzamanM.ChaudharyS.AzeemI.et al. (2024). Nanotechnology in precision agriculture: Advancing towards sustainable crop production. Plant Physiol. Biochem.206, 108244. doi: 10.1016/j.plaphy.2023.108244

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 95

  ZhaiW.ChengQ.DuanF.HuangX.ChenZ. (2025). Remote sensing-based analysis of yield and water-fertilizer use efficiency in winter wheat management. Agric. Water Manage.311, 109390. doi: 10.1016/j.agwat.2025.109390

  • CrossRef
  • Google Scholar

• 96

  ZhaiZ.MartínezJ. F.BeltranV.MartínezN. L. (2020). Decision support systems for agriculture 4.0: Survey and challenges. Comput. Electron. Agric.170, 105256. doi: 10.1016/j.compag.2020.105256

  • CrossRef
  • Google Scholar

• 97

  ZhaoW.QuY.ChenJ.YuanZ. (2020). Deeply synergistic optical and SAR time series for crop dynamic monitoring. Remote Sens. Environ.247, 111952. doi: 10.1016/j.rse.2020.111952

  • CrossRef
  • Google Scholar

• 98

  ZhaoW.WangM.PhamV. T. (2023). Unmanned aerial vehicle and geospatial analysis in smart irrigation and crop monitoring on ioT platform. Mobile Inf. Syst.2023, 4213645. doi: 10.1155/2023/4213645

  • CrossRef
  • Google Scholar

• 99

  ZhuL.WuF.FuK.HuY.WangY.TianX.et al. (2023). An active service recommendation model for multi-source remote sensing information using fusion of attention and multi-perspective. Remote Sens.15, 2564. doi: 10.3390/rs15102564

  • CrossRef
  • Google Scholar

• 100

  ZualkernanI.AbuhaniD. A.HussainM. H.KhanJ.ElMohandesM. (2023). Machine learning for precision agriculture using imagery from unmanned aerial vehicles (UAVs): A survey. Drones7, 382. doi: 10.3390/drones7060382

  • CrossRef
  • Google Scholar