Using AI Chat and Satellite Maps for Farm Decisions: A Practical Evaluation of Available Tools

Introduction

Over the past few decades, our "eyes in the sky" have increased in number and monitoring capabilities. Satellite constellations such as Landsat, Sentinel, and MODIS can track field-scale ecosystem information, including soil conditions, water dynamics, and shifts in crop health (Gorelick et al. 2017; Wulder et al. 2019; Gallios et al. 2026). These data are commonly accessed through web-based platforms and dashboards that display satellite-derived maps and indicators. At the same time, advances in artificial intelligence (AI) have made it much faster and easier to process this information and find useful patterns (Reichstein et al. 2019; Ahmad and Jhara 2024). This supports the development of products that can further improve agro-environmental monitoring and relevant applications.

Nevertheless, many farmers and land managers still find these tools difficult to use (Pinna et al. 2024). Specifically, many platforms and dashboards require account registration, payment, technical skills, data uploads, or specialized software, which all limit the practical adoption of these technologies in agriculture (Rose and Chilvers 2018). Additionally, the processing and delivery of actionable information from these platforms may take too long to fit within the narrow decision windows of farm operations (e.g., irrigation, fertilization) and post-event responses (Wolfert et al. 2017). Thus, valuable data that could support agricultural needs and sustainable farming practices often remain underutilized.

A promising way to close this gap is by using geospatial AI (GeoAI) agents to ask questions in everyday language and receive clear, actionable information without manual data processing (Zhai et al. 2020). However, the rapid emergence of numerous AI-based geospatial tools has created a new challenge for agricultural stakeholders. Specifically, a growing number of platforms now claim to provide satellite-based insights through conversational interfaces or AI assistants, yet their capabilities, reliability, and accessibility vary widely. For farmers, consultants, and Extension professionals who already face time and resource constraints, it can be difficult to determine which tools are useful, trustworthy, or practical for real-world farm management. As a result, there is a need for a clear and structured evaluation of what the currently available tools can deliver for agricultural decision-making.

Given this context, the purpose of this publication is to evaluate currently available GeoAI agents using a consistent set of inclusion criteria and agricultural prompts. The goal is to provide farmers, consultants, and Extension agents with a practical assessment of what these tools can and cannot do today, highlight current limitations, and discuss their potential role in supporting future agricultural decision-making. Please note that this publication is not intended to propose a new evaluation framework, but rather to provide a practical, use-oriented assessment of currently available tools and their relevance for real-world agricultural applications.

Tool Inspection via the Inclusion Rule

Considering the vast number of geospatial tools available today, this publication focuses on identifying platforms that could support farmers as AI-based agents and that combine maps, data, and interactive guidance in one place. More precisely, such promising tools are helpful with geospatial processing and ensure that farmers get field-specific insights rather than general information. AI chat (LLM integration) increases the probability of farmers adopting digital agricultural tools when they are easy to use and provide guidance in everyday language (Marshall et al. 2021; Beke et al. 2024). Moreover, ease of access, minimal registration, and low technical barriers strongly influence adoption and continued use by farmers (Sishodia et al. 2020; Metternicht et al. 2022; Marshall et al. 2022). Considering this information, an initial structured search was conducted using online sources, scientific literature, and product documentation to identify commonly accessible platforms. This search was intended as a general exploration rather than a formal systematic review. To identify relevant tools, general keyword searches were conducted using online search engines (e.g., Google), product pages, and user documentation. Search keywords focused on GeoAI and agricultural decision-support tools, for example, “geospatial AI for agriculture” and “chat-based mapping tools.”

Based on this process, a simple inclusion rule was applied to select tools that could function as GeoAI agents in an agricultural context. A platform was included only if it met all three of the following conditions:

1. Geospatial processing: The tool must be able to process and analyze geospatial and satellite data for agricultural fields.
2. Conversational interaction: The platform must allow users to ask questions in plain language and receive explanations.
3. Accessibility: The tool must be free of cost or offer a free trial.

In other terms, a platform was included only if it met all three of the following conditions:

Inclusion rule for platform selection = (Geospatial processing capability) AND (conversational interaction) AND (free or trial access).

Please note that because the second criterion required chat capabilities, only platforms that enable farmers to ask questions and receive plain-language explanations were considered. In this regard, many traditional geospatial dashboards and platforms, both freely accessible and commercial, have been excluded. In particular, traditional dashboards and platforms that are free of charge or provide limited access (OneSoil, Planet Insights Platform, Sentinel Hub, and Descartes Labs), as well as those for purely commercial use (IBM Environmental Intelligence Suite, Tend, Cropin, Climate FieldView, Taranis, Terensis, Agronaut, Farmonaut, GeoTerra, Carto, and AI Superior), offer powerful capabilities for satellite data visualization and spatial analytics without supporting conversational interaction using AI. Therefore, users must independently interpret maps, indices, and charts, which can limit accessibility for non-technical audiences and increase the time required to translate data into management decisions. On the other hand, several platforms support AI-based chat interaction but do not offer spatial analytics and visualizations. AI chat-based systems such as ChatGPT, DeepSeek, Claude, Gemini, and the AskIFAS ExtensionBot meet the conversational interaction criterion but lack direct connection with data and geospatial processing capabilities; hence, they have been excluded from this review. This approach prioritizes ease of use, access, and practical functionality, which strongly influences whether tools are adopted by farmers, consultants, and Extension professionals.

Following the application of the inclusion rule, 13 platforms were identified as potential candidate GeoAI agents: Spatchat, Gaia Bot, EOSDA Crop monitoring, ASKTERRA, Esri ArcGIS Assistant (ArcGIS Copilot), Google Earth AI (Gemini integration), Google Earth Engine + Gemini, Navi (by Atlas), Manteo.ai, KissanAI, CropX, Xsupra, and FarmHawk (Beta). However, some of these platforms were excluded because they required longer approval times (more than five working days), only operated in specific regions, or were very technical/required specialized software, which reduced their likelihood of adoption in an agricultural setting. These accessibility constraints excluded the following platforms: Esri ArcGIS Assistant (ArcGIS Copilot), Google Earth AI (Gemini integration), Google Earth Engine + Gemini, Navi (by Atlas), Manteo.ai, KissanAI, and CropX. In addition, Spatchat, ASKTERRA, Xsupra, and FarmHawk (Beta) met the inclusion rule and further accessibility constraints and were thus selected for testing. However, Spatchat, ASKTERRA, and FarmHawk (Beta) could not be fully evaluated due to technical errors that prevented the generation of complete evaluable outputs. Furthermore, Xsupra could not be fully evaluated because some relevant functionality required payment. To reiterate, 13 platforms met the inclusion criteria. Of these, seven were excluded from testing due to access, regional, or technical-use limitations. The remaining six were selected for testing. Among the six tested platforms, three could not be fully evaluated due to technical errors, one could only be partially assessed, and two were successfully assessed.

For each of the identified GeoAI agents, Table 1 reports its access status (open/research, commercial, or limited access [commercial with free trial]), primary applications, type of chat and map support, and whether it was tested in this evaluation. It also includes lists of intended primary users because most selected GeoAI agents are not designed primarily for farmers but also target businesses, researchers, educators, consultants, or broader environmental and spatial specialists.

Comparison Criteria and Testing

Once a group of GeoAI agents was selected and filtered, relevant peer-reviewed literature was consulted to identify the key factors influencing farmers’ adoption and effective use of digital agricultural technologies. These factors informed the criteria used to compare platforms (Figure 1) (McCormack et al. 2021). More precisely, research shows that practical accessibility and setup requirements (e.g., registration, data input requirements, and ease of field selection), platform access status (open/research vs. commercial/limited), and intended primary users influence adoption. These design objectives contribute to perceived usefulness, ease of use, and compatibility with farm workflows (Thomas et al. 2023). In addition, practical usefulness further depends on functionality, such as spatial analysis, forecasting, modeling tools, and tailored recommendations, aligned with farmers’ decision-making processes and local conditions (Lundström and Lindblom 2018; McCormack et al. 2021; Beke et al. 2024). Likewise, transparency in data sources and response time are strongly associated with reliability, user satisfaction, and sustained engagement (Wang et al. 2023, 2024).

Thereafter, three prompts were developed in consultation with agricultural experts to test the practicality of these platforms:

1. “I’m getting inconsistent germination across my green bean field. Could this be related to differences in soil moisture or drainage?”
2. “We had a hard freeze in late January that may have damaged my cucumbers. Can you show whether some fields were affected more than others?”
3. “What is the soil pH across my field?”

For tests requiring spatial data as input from the user, polygons representing one or two experimental fields were created at the UF/IFAS Southwest Florida Research and Education Center in Immokalee, Florida (SWFREC; N 26°27′41.8″, W 81°26′47.5″). In some cases, to test the functionality of these tools, satellite-derived vegetation index (Normalized Difference Vegetation Index [NDVI]) layers were created to cover the same fields and buffered areas (Figure 2). More precisely, this index depicts values ranging from approximately 0.4 (red hue) to 0.76 (green hue), indicating vegetation cover. Higher values illustrate healthier vegetation, whereas lower values suggest potential crop stress within the field (without further information on the causal factor). In this publication, the NDVI layer was generated through the preprocessing and processing of Sentinel-2 surface reflectance data using the Google Earth Engine platform, providing a representative snapshot of field conditions. Additionally, the NDVI layer was used as contextual spatial input to support platform interaction and testing, rather than as ground truth for evaluating platform accuracy. Accordingly, this publication does not aim to validate or quantify the accuracy of GeoAI outputs against reference data, but rather to assess the usability, responsiveness, and practical relevance of the platforms under consistent input conditions. Finally, if a platform did not show where its information came from, a follow-up question was posed to see whether it could provide the data sources.

Selected Geospatial AI Agents for Agricultural Recommendations

Although 13 platforms satisfied the inclusion rule, not all of them could be evaluated due to practical limitations during the review process. Additional constraints referred to existing accessibility limitations of GeoAI agents, leading to the exclusion of those that did not follow this convention:

Platforms had to be accessible within five working days and operational for the region of interest, and they should not require specialized knowledge or advanced technical setup.

This additional filter depicts probable adoption conditions in agriculture, where extended approval processes and complex platforms can limit usability. Table 2 presents the subset of GeoAI agents that met the inclusion rule criteria, satisfied the practical accessibility constraints, and were, therefore, tested using the agricultural prompts. This distinction clarifies that exclusion from the subset in Table 2 does not imply conceptual inadequacy, but rather limited practical accessibility. Based on the comparison criteria from Figure 1, the outputs of the analysis were classified into four categories: “High,” “Medium,” “Low,” and “Not confirmed." The first three categories reflect the level of performance or practical suitability of each GeoAI platform, whereas “not confirmed” was used when technical errors prevented the platform from being evaluated.

In addition, Figures 3–6 complement Table 2 by visually illustrating examples of the performance and outputs of the tested platforms as each GeoAI agent handled agricultural prompts.

Spatchat is a free, web-based tool that does not require registration and is relatively easy to use. It requires users to upload raster/image (TIF [Tag Image File Format]) data and supports modules for landscape metrics and basic statistics, which are relevant to agricultural land evaluation. The platform accepts plain-language requests and is designed to produce maps, metrics, and written explanations. During its testing, the uploaded image file returned only basic metadata (e.g., extent, coordinate system, pixel size), with response times ranging from 3 to 16.5 seconds. However, no maps, charts, or calculated metrics were generated due to possible technical errors. Therefore, its ability to compute landscape metrics and statistical summaries could not be confirmed. Overall, the tool shows potential but, in this test, was limited to reporting file metadata rather than producing landscape analysis outputs.

Another tested platform was Gaia Bot, a free, web-based tool developed by the UF/IFAS Soil AI Lab at the SWFREC. It does not require registration or user-uploaded spatial data (Gallios et al. 2025). Instead, users define their area of interest by drawing a polygon on the map, which is practical for producers and agencies. Then, the platform provides written explanations, visual outputs, and site-specific recommendations, including information on vegetation health and key soil properties (e.g., pH and soil organic carbon), as well as irrigation and fertilization recommendations. It also asks follow-up questions to support further analysis. Moreover, Gaia Bot responded to two out of the three agricultural prompts within 3 to 18 seconds. Outputs were generally relevant to the questions and presented in a clear text-and-map format suitable for nontechnical users. However, the platform does not provide citations or data sources for its recommendations, which may limit its use.

In addition, EOSDA Crop Monitoring is a commercial platform (free trial provided) that requires user registration but does not require users to upload their own spatial datasets. Users define fields of interest, after which the system provides access to remote sensing–derived products related to vegetation and moisture over extended time periods. In addition, the platform includes a chat feature that generates explanations and general guidance about its products. However, it does not provide detailed text for interpreting the data, and outputs are limited to basic statistical summaries and simple classifications (e.g., vegetation data categories [bare soil, sparse vegetation, and dense vegetation]). The interface is easy to use, and the primary outputs consist of visualizations and basic statistics that support visual inspection of crop conditions and moisture patterns.

Furthermore, Xsupra is a commercial GeoAI agent that offers a free trial and requires simple, quick registration. Users do not need to upload their own datasets; instead, they can define their fields by drawing an area of interest directly on the map. The platform also provides an option to upload spatial data for visualization and analysis, allowing users to inspect datasets that may not already be available within the system. In addition, the platform includes several vegetation indices and meteorological data layers that help depict field conditions, while the chat interface can provide explanations and relevant field characteristics. During testing, the platform generated outputs for Agricultural Prompts 1 and 3, providing general guidance (non-region specific), but Agricultural Prompt 2 could not be evaluated because it required access through a paid plan. According to the product description, the platform can generate chat responses, visualizations, and site-specific guidance for agricultural decision-making. However, it could not be entirely tested.

Lastly, ASKTERRA and FarmHawk (Beta) are commercial platforms (free trial provided) similar to Gaia Bot that provide data, visualizations, and written explanations in response to the user’s questions. However, during testing, the free plan did not produce results for any of the prompts (no response). As such, their usability and functionality could not be evaluated and confirmed in this publication.

Notably, response times for agricultural prompts ranged from approximately 3 to 11 seconds (except for Xsupra, which took 37 to 50 seconds). The answers were relevant to the questions asked, although no sources of references/citations were provided. The EOSDA Crop Monitoring platform offered general guidance for all three agriculture-related questions, but it was not site-specific and relied mainly on general knowledge. Overall, the EOSDA Crop Monitoring platform is useful for monitoring crop conditions over time but provides limited analytical depth and decision-making support beyond general observations and indices.

Key Findings

• Accessibility and initiation: Three of the evaluated GeoAI agents are free of charge with no limitations or paid plans, four require user registration, and only one (Spatchat) requires user-provided data. The free cost (or free trial) and minimal setup reduce barriers to adoption, allowing farmers to test tools quickly without investing time, money, or technical effort.
• Site-specific guidance: Among the successfully tested GeoAI agents, only Gaia Bot provides field-level management guidance. This type of localized output is particularly valuable for farm decisions because it links satellite data to specific management zones.
• Response time: All GeoAI agents, except Xsupra, delivered answers within approximately 3 to 20 seconds (using a high-speed internet connection), making them suitable for rapid information retrieval. This is especially important in agriculture, where many management decisions must be made within narrow time windows.
• Source transparency: None of the evaluated GeoAI agents provided citations, even when requested, which may reduce transparency in their recommendations. This limitation should be addressed in existing and forthcoming platforms for further adoption and usability of these tools.

Collectively, these findings suggest that GeoAI agents currently perform best as rapid visualization and general agronomic guidance tools rather than comprehensive site-specific decision-support systems.

Practical Uses for Producers and Agencies

Based on the evaluation conducted in this publication, the following questions and recommendations can help different stakeholders make effective use of GeoAI agents:

• When are GeoAI tools useful?
• For rapid field assessment, visualization, and general crop monitoring
• For exploring “what-if” scenarios and supporting preliminary decision-making
• For generating maps and summaries that complement field observations
• When should one use caution?
• For high-stakes management decisions (e.g., fertilizer rates, pesticide applications, major investments)
• When platform outputs lack transparency, sources, or clear data provenance
• In situations where local conditions (soil, management history, weather) are not fully captured
• What are some safe practices for use?
• Always verify outputs with field observations, local knowledge, and trusted data sources
• Ask for data sources and explanations when available; be cautious if none are provided
• Use multiple tools or data sources, when possible, to cross-check results.

These tools can support farmers with day-to-day field insights. Consultants may use them for rapid assessments and scenario exploration, and Extension professionals can leverage them for education, demonstration, and decision-support training. However, they should be used with caution and close attention to the guidance presented in this publication.

Practical Limitations for Producers and Agencies

Despite rapid advancements in the geospatial and artificial intelligence worlds, based on the tools assessed here, no single conversational platform currently provides complete decision support or recommendations that are site-specific, easy to use, and supported by verified sources. In practice, this means that while GeoAI agents can quickly summarize farm conditions or highlight potential vegetation stress, their recommendations often rely on generalized agronomic knowledge rather than field-level variables such as cultivar selection, soil type, drainage conditions, irrigation practices, and management history. As a result, responses may appear confident and technically coherent but may not fully reflect local production realities, potentially leading to inappropriate or oversimplified recommendations.

Additional limitations relate to platform and market stability, as several of the reviewed systems operate under limited-access, beta, or evolving commercial models. Consequently, features, pricing structures, model capabilities, and availability may change rapidly overtime. Moreover, the results presented in this publication are based on a limited number of structured prompts, and testing occurred only within a specific geographic and production context (SWFREC), using a high-speed internet connection. However, in real terms, platform performance may vary substantially depending on region, crop type, seasonal conditions, data availability, and user skill.

Finally, input validation across platforms was generally weak, as none of the evaluated GeoAI agents detected incorrect or incomplete user inputs during testing. For example, when vegetation data from only one field were uploaded in response to a comparative question requiring at least two farms (Agricultural Prompt 2), the systems still generate confident but incomplete or erroneous responses. This limitation highlights a broader concern: GeoAI agents may not reliably flag missing data, logical inconsistencies, or improper analytical setups. Users must therefore critically assess outputs and verify that inputs are appropriate before acting on AI-generated recommendations.

Conclusions

In a nutshell, GeoAI agents can change how growers and Extension agents interact with satellite and land data by turning complex maps and remote sensing analytics into simple question-and-answer guidance. In this publication, a total of 13 GeoAI agents met the inclusion rule criteria of being useful (field-specific geospatial analytics), usable (plain-language interaction with explanations), and adoptable (low cost and low setup burden in real farm workflows). However, seven of them were excluded from testing because of further accessibility, location, or technical-use constraints. Of the remaining six, two (Gaia Bot and EOSDA Crop Monitoring) were fully assessed, one (Xsupra) was partially assessed, and three could not be fully evaluated because of technical errors. Moreover, response times were very fast for all evaluated agents, ranging from 3 to 20 seconds (except for Xsupra, which took 37 to 50 seconds), while only one agent (Gaia Bot) provided field-level guidance. Nevertheless, none of the evaluated GeoAI agents incorporated all essential criteria (as defined in Figure 1), including valid sources of information, within a single platform. Therefore, existing and future GeoAI agents should go beyond basic chat and geospatial functions. They should follow a hybrid workflow in which their existing capabilities are complemented by evidence-based recommendations that combine near real-time satellite data with local and scientific knowledge. From a practical perspective, GeoAI agents can be useful for rapid field scouting, visualization, and preliminary assessments of crop conditions. However, they should not replace professional judgment or field-based observations. Instead, their outputs should be interpreted in combination with local knowledge, management history, and environmental conditions. In addition, users are encouraged to verify the sources of information provided by these platforms and to cross-check results, when possible, especially for decisions with economic or environmental implications. By emphasizing accessibility, ease of use, locally relevant information, and transparency, GeoAI agents can transform “data overload” into meaningful conversations that help farmers, land managers, and Extension agents make more informed and sustainable decisions.

Acknowledgments

This work was supported by project award no. 2025-68016-44493 from the U.S. Department of Agriculture, National Institute of Food and Agriculture (USDA NIFA).

References

Ahmad, H., and S. I. Jhara. 2024. “AI-Driven Approaches for Real-Time Satellite Data Processing and Analysis.” PowerPoint Presentation, NASA Accelerating Informatics for Earth Science, Arlington, VA, June 13, 2024. https://assets.science.nasa.gov/content/dam/science/cds/science-enabling-technology/events/2025/accelerating-informatics/PM_6_Ahmad.pdf

Beke, B., M. S. Segovia, and C. Valdivia. 2024. “Adoption of Smart Farm Networks: A Translational Process to Inform Digital Agricultural Technologies.” Agriculture and Human Values 41: 1573–1590. https://doi.org/10.1007/s10460-024-10566-3

Gallios, G., J. A. M. Demattê, N. Tsakiridis, M. C. Cardoso, A. Kritharoula, and N. Tziolas. 2026. “Federated Earth-Observation Models for Collaborative Farm-Scale Soil Mapping.” International Journal of Applied Earth Observation and Geoinformation 146: 105067. https://doi.org/10.1016/j.jag.2025.105067

Gallios, G., A. Kritharoula, and N. Tziolas. 2025. “GaiaBot: Simplifying Access to Soil Data.” In IGARSS 2025—2025 IEEE International Geoscience and Remote Sensing Symposium. https://doi.org/10.1109/igarss55030.2025.11242316

Gorelick, N., M. Hancher, M. Dixon, S. Ilyushchenko, D. Thau, and R. Moore. 2017. “Google Earth Engine: Planetary-Scale Geospatial Analysis for Everyone.” Remote Sensing of Environment 202: 18–27. https://doi.org/10.1016/j.rse.2017.06.031

Lundström, C., and J. Lindblom. 2018. “Considering Farmers’ Situated Knowledge of Using Agricultural Decision Support Systems (AgriDSS) to Foster Farming Practices: The Case of CropSAT.” Agricultural Systems 159: 9–20. https://doi.org/10.1016/j.agsy.2017.10.004

Marshall, A., K. Turner, C. Richards, M. Foth, and M. Dezuanni. 2021. “Rural Data Divides: Critical Factors of Digital Agtech Adoption on Australian Farms.” AoIR Selected Papers of Internet Research. https://doi.org/10.5210/spir.v2021i0.12206

Marshall, A., K. Turner, C. Richards, M. Foth, and M. Dezuanni. 2022. “Critical Factors of Digital AgTech Adoption on Australian Farms: From Digital to Data Divide.” Information, Communication & Society 25 (6): 868–886. https://doi.org/10.1080/1369118X.2022.2056712

McCormack, M., C. Buckley, and E. Kelly. 2021. “Using a Technology Acceptance Model to Investigate What Factors Influence Farmer Adoption of a Nutrient Management Plan.” Irish Journal of Agricultural and Food Research 60 (1). https://doi.org/10.15212/ijafr-2020-0134

Metternicht, G., R. Cross, N. Perry, et al. 2022. “Opportunities and Barriers for Adoption of Sentinel-2 Based Agricultural Services and Products: Policy and Stakeholders’ Views as Influencers.” In IGARSS 2022—2022 IEEE International Geoscience and Remote Sensing Symposium. https://doi.org/10.1109/IGARSS46834.2022.9883853

Pinna, D., A. Pezzuolo, A. Cogato, C. Pornaro, S. Macolino, and F. Marinello. 2024. “Applications of Satellite Platforms and Machine Learning for Mapping and Monitoring Grasslands and Pastures: A Systematic and Comprehensive Review.” Smart Agricultural Technology 9: 100571. https://doi.org/10.1016/j.atech.2024.100571

Reichstein, M., G. Camps-Valls, B. Stevens, et al. 2019. “Deep Learning and Process Understanding for Data-Driven Earth System Science.” Nature 566 (7743): 195–204. https://doi.org/10.1038/s41586-019-0912-1

Rose, D. C., and J. Chilvers. 2018. “Agriculture 4.0: Broadening Responsible Innovation in an Era of Smart Farming.” Frontiers in Sustainable Food Systems 2: 87. https://doi.org/10.3389/fsufs.2018.00087

Sishodia, R. P., R. L. Ray, and S. K. Singh. 2020. “Applications of Remote Sensing in Precision Agriculture: A Review.” Remote Sensing 12 (19): 3136. https://doi.org/10.3390/rs12193136

Thomas, R. J., G. O’Hare, and D. Coyle. 2023. “Understanding Technology Acceptance in Smart Agriculture: A Systematic Review of Empirical Research in Crop Production.” Technological Forecasting and Social Change 189: 122374. https://doi.org/10.1016/j.techfore.2023.122374

Wang, S., X. Huang, P. Liu, et al. 2024. “Mapping the Landscape and Roadmap of Geospatial Artificial Intelligence (GeoAI) in Quantitative Human Geography: An Extensive Systematic Review.” International Journal of Applied Earth Observation and Geoinformation 128: 103734. https://doi.org/10.1016/j.jag.2024.103734

Wang, J., S. Zhang, B. Liu, and L. Zhang. 2023. "Decision Making with the Use of Digital Inclusive Financial Systems by New Agricultural Management Entities in Guangdong Province, China: A Unified Theory of Acceptance and Use of Technology-Based Structural Equation Modeling Analysis.” Systems 11 (10): 513. https://doi.org/10.3390/systems11100513

Wolfert, S., L. Ge, C. Verdouw, and M.-J. Bogaardt. 2017. “Big Data in Smart Farming—A Review.” Agricultural Systems 153: 69–80. https://doi.org/10.1016/j.agsy.2017.01.023

Wulder, M. A., T. R. Loveland, D. P. Roy, et al. 2019. “Current Status of Landsat Program, Science, and Applications.” Remote Sensing of Environment 225: 127–147. https://doi.org/10.1016/j.rse.2019.02.015

Zhai, Z., J. F. Martínez, V. Beltran, and N. L. Martínez. 2020. “Decision Support Systems for Agriculture 4.0: Survey and Challenges.” Computers and Electronics in Agriculture 170: 105256. https://doi.org/10.1016/j.compag.2020.105256