Agricultural multimodal reasoning requires robust spatial understanding across varying scales, from ground-level close-ups to top-down UAV and satellite imagery. Existing Multi-modal Large Language Models (MLLMs) suffer from a significant "terrestrial-centric" bias, causing scale confusion and logic drift during complex agricultural planning. To address this, we introduce the first large-scale AgroOmni (288K), a multi-view training corpus designed to capture diverse spatial topologies and scales in modern precision agriculture. Built on this dataset, we propose AgroNVILA, an MLLM that utilizes a novel Perception-Reasoning Decoupling (PRD) architecture. On the perception side, we incorporate a View-Conditioned Meta-Net (VCMN), which injects macroscopic spatial context into visual tokens, resolving scale ambiguities with minimal computational overhead. On the reasoning side, Agriculture-aware Relative Policy Optimization (ARPO) leverages reinforcement learning to align the model's decision-making with expert agricultural logic, preventing statistical shortcuts. Extensive experiments demonstrate that AgroNVILA outperforms state-of-the-art MLLMs, achieving significant improvements (+15.18%) in multi-altitude agricultural reasoning, reflecting its robust capability for holistic agricultural spatial planning.
YANG Yunxia, LU Jun, SONG Xulei, WANG Xin, WANG Heng, WANG Deliang, YANG Yigong, ZHONG Chonghu, ZHANG Fuyu
In order to study the effect of different distillation methods on the flavor of new whisky, selecting two common distillation methods (pot distillation and tower distillation) in this study, the initial distillate of the first distillation was distilled twice, and the sample was taken in stages and then evaluated by sensory evaluation. The volatile flavor substances of 22 distilled spirits samples were determined by gas chromatography (GC), and the effects of two different secondary distillation methods on the flavor of new whisky were compared. The results showed that the flavor of new whiskey distilled by the two distillation methods was different. The new whiskey distilled by pot distillation was biased towards floral aroma, while the new whiskey distilled by tower distillation was more biased towards fruity and sweet aroma. A total of 40 volatile flavor compounds were detected in new whisky, of which aldehydes and esters were mainly in the head part the wine, and esters were mainly ethyl acetate. Among the alcohols, the content of methanol in the core part of the whisky was low, and high in the tail part of the whisky. The volatile flavor substances in the new whiskey distilled by pot distillation were mainly alcohols, while the volatile flavor substances in the new whiskey distilled by tower distillation were mainly aldehydes and esters.
Real-time assessment of trunk growth is vital for understanding tree growth fluctuation, especially under irrigation application and other environmental factors. Accurate trunk diameter assessment is crucial for optimizing water use and tree health improvement, and its cost-effectiveness is needed for widespread adoption in agriculture. This study focused on the development of an accurate and low-cost IoT-based dendrometer system for real-time trunk diameter measurement of Christmas trees. The dendrometer sensor was calibrated (R2 = 0.99) to ensure the accurate conversion of sensor voltage to trunk diameter fluctuations. This IoT-based dendrometer system consists of a platform that enables wireless data transmission, cloud-based storage and real-time analysis of the trunk diameter. Temperature fluctuation influenced the sensor readings with no impact, which validated the system's reliability in open field conditions. Christmas tree diameter monitoring showed significant trunk expansion and contraction under irrigation application and water stress, respectively, which signifies the system ability to monitor the real-time trunk growth responses. Cost analysis makes this technology economical and reliable for widespread application in precision agriculture. Therefore, this low-cost IoT-based dendrometer system is reliable, accurate, and economically viable for improving irrigation management, tree health monitoring, and supporting farmers through data-driven agricultural practices.
Abstract Background The global burden of early-onset chronic diseases, especially early-onset type 2 diabetes, is increasing, particularly in China. Diet is a key factor and emerging evidence highlights substantial inter-individual variability in metabolic responses to diets, highlighting the need for precision nutrition. Methods The China Precision Nutrition Biobank (CPNB) is a prospective, longitudinal, cohort study designed to investigate diet-phenotype/genotype interactions and develop precision dietary strategies for early prevention and intervention of chronic diseases, with a particular focus on early-onset diseases. CPNB consists of three phases: the alpha (pilot cohort), beta (transition cohort), and gamma (main cohort) phases. Approximately 200, 1450, and 20,000 adults aged 18–40 years from urban and rural areas in China including Beijing, one city each in Heilongjiang, Shandong, Zhejiang, Guangxi, and Hainan provinces, and one or more villages each in Henan, Gansu, Sichuan, Zhejiang, and Hunan provinces will be recruited during the alpha, beta, and gamma phases, respectively, between 2025 and 2035. Sociodemographic information, medical records, read-time weighed food records and corresponding continuous glucose monitor (CGM) readings, objective physical activity, food challenges, genes, gut and oral microbiota, metabolites from blood, stool, urine, saliva, and hair, and questionnaires will be collected at baseline survey. The follow-up survey will be conducted every five years to repeat these assessments until participants’ death (the follow-up period may extend up to 80 years). Outcomes of interest are common early- and late-onset chronic diseases and their preclinical stages. Discussion The CPNB data can be used to develop prediction models for personalized metabolic responses and risks of early-onset chronic diseases among Chinese people. It will also provide new evidence on interactions of diet with human phenotypes/genotypes during preclinical stage, onset, and progression of early-onset diseases. CPNB aims to inform the development of precision nutrition strategies aligned with the principles of predictive, personalized, preventive, and participatory medicine in the Chinese population. Trial registration CPNB was registered at Chinese Clinical Trial Registry ( http://www.chictr.org.cn/ ) on June 3rd, 2025, under the registration number ChiCTR2500103621.
Nutrition. Foods and food supply, Nutritional diseases. Deficiency diseases
Vegetation is a fundamental component of terrestrial ecosystems, and accurately assessing the effects of seasonal climate variations, extreme weather events, and land use changes on vegetation dynamics is crucial. The Ganjiang River Basin (GRB), a key region for water conservation and recharge in southeastern China, has experienced significant land use changes and variable climate in the past. However, comprehensive evaluations of how these changes have impacted vegetation remain limited. To address this gap, we used machine learning models (random forest and XGBoost) to assess the impact of seasonal and extreme climate variables, land cover, topography, soil properties, atmospheric CO<sub>2</sub>, and night-time light intensity on vegetation dynamics. We found that the annual mean NDVI showed a slight increase from 1990 to 1999 but has decreased significantly over the last 8 years. XGBoost was better than the RF model in simulating the NDVI when using all five types of data source (R<sup>2</sup> = 0.85; RMSE = 0.04). The most critical factors influencing the NDVI were forest and cropland ratio, followed by soil organic carbon content, elevation, cation exchange capacity, night-time light intensity, and CO<sub>2</sub> concentration. Spring minimum temperature was the most important seasonal climate variable. Both linear and nonlinear relationships were identified between these variables and the NDVI, with most variables exhibiting threshold effects. These findings underscore the need to develop and implement effective land management strategies to enhance vegetation health and promote ecological balance in the region.
With the wide application of computer vision in agriculture, image analysis has become the key to tasks such as crop health monitoring and pest detection. However, the significant domain shifts caused by environmental changes, different crop types, and diverse data acquisition methods seriously hinder the generalization ability of the model in cross-region, cross-season, and complex agricultural scenarios. This paper explores how domain adaptation (DA) techniques can address these challenges to improve cross-domain transferability in agricultural image analysis. DA is considered a promising solution in the case of limited labeled data, insufficient model adaptability, and dynamic changes in the field environment. This paper systematically reviews the latest advances in DA in agricultural images in recent years, focusing on application scenarios such as crop health monitoring, pest and disease detection, and fruit identification, in which DA methods have significantly improved cross-domain performance. We categorize DA methods into shallow learning and deep learning methods, including supervised, semi-supervised and unsupervised strategies, and pay special attention to the adversarial learning-based techniques that perform well in complex scenarios. In addition, this paper also reviews the main public datasets of agricultural images, and evaluates their advantages and limitations in DA research. Overall, this study provides a complete framework and some key insights that can be used as a reference for the research and development of domain adaptation methods in future agricultural vision tasks.
Agricultural robotic agents have been becoming useful helpers in a wide range of agricultural tasks. However, they still heavily rely on manual operations or fixed railways for movement. To address this limitation, the AgriVLN method and the A2A benchmark pioneeringly extend Vision-and-Language Navigation (VLN) to the agricultural domain, enabling agents to navigate to the target positions following the natural language instructions. We observe that AgriVLN can effectively understands the simple instructions, but often misunderstands the complex ones. To bridge this gap, we propose the T-araVLN method, in which we build the instruction translator module to translate noisy and mistaken instructions into refined and precise representations. When evaluated on A2A, our T-araVLN successfully improves Success Rate (SR) from 0.47 to 0.63 and reduces Navigation Error (NE) from 2.91m to 2.28m, demonstrating the state-of-the-art performance in the agricultural VLN domain. Code: https://github.com/AlexTraveling/T-araVLN.
Dinesh Jackson Samuel, Inna Skarga-Bandurova, David Sikolia
et al.
AgroLLM is an AI-powered chatbot designed to enhance knowledge-sharing and education in agriculture using Large Language Models (LLMs) and a Retrieval-Augmented Generation (RAG) framework. By using a comprehensive open-source agricultural database, AgroLLM provides accurate, contextually relevant responses while reducing incorrect information retrieval. The system utilizes the FAISS vector database for efficient similarity searches, ensuring rapid access to agricultural knowledge. A comparative study of three advanced models: Gemini 1.5 Flash, ChatGPT-4o Mini, and Mistral-7B-Instruct-v0.2 was conducted to evaluate performance across four key agricultural domains: Agriculture and Life Sciences, Agricultural Management, Agriculture and Forestry, and Agriculture Business. Key evaluation metrics included embedding quality, search efficiency, and response relevance. Results indicated that ChatGPT-4o Mini with RAG achieved the highest accuracy at 93%. Continuous feedback mechanisms enhance response quality, making AgroLLM a benchmark AI-driven educational tool for farmers, researchers, and professionals, promoting informed decision-making and improved agricultural practices.
The large-scale integration of robots in agriculture offers many promises for enhancing sustainability and increasing food production. The numerous applications of agricultural robots rely on the transmission of data via mobile network, with the amount of data depending on the services offered by the robots and the level of on-board technology. Nevertheless, infrastructure required to deploy these robots, as well as the related energy and environmental consequences, appear overlooked in the digital agriculture literature. In this study, we propose a method for assessing the additional energy consumption and carbon footprint induced by a large-scale deployment of agricultural robots. Our method also estimates the share of agricultural area that can be managed by the deployed robots with respect to network infrastructure constraints. We have applied this method to metropolitan France mobile network and agricultural parcels for five different robotic scenarios. Our results show that increasing the robot's bitrate needs leads to significant additional impacts, which increase at a pace that is poorly captured by classical linear extrapolation methods. When constraining the network to the existing sites, increased bitrate needs also comes with a rapidly decreasing manageable agricultural area.
With the global population increasing and arable land resources becoming increasingly limited, smart and precision agriculture have emerged as essential directions for sustainable agricultural development. Artificial intelligence (AI), particularly deep learning models, has been widely adopted in applications such as crop monitoring, pest detection, and yield prediction. Among recent generative models, diffusion models have demonstrated considerable potential in agricultural image processing, data augmentation, and remote sensing analysis. Compared to traditional generative adversarial networks (GANs), diffusion models exhibit greater training stability and superior image generation quality, effectively addressing challenges such as limited annotated datasets and imbalanced sample distributions in agricultural scenarios. This paper reviews recent advancements in the application of diffusion models within agriculture, focusing on their roles in crop disease and pest detection, remote sensing image enhancement, crop growth prediction, and agricultural resource management. Diffusion models have been found useful in improving tasks like image generation, denoising, and data augmentation in agriculture, especially when environmental noise or variability is present. While their high computational requirements and limited generalizability across domains remain concerns, the approach is gradually proving effective in real-world applications such as precision crop monitoring. As research progresses, these models may help support sustainable agriculture and address emerging challenges in food systems.
In this study, the impact of research and development (R&D) expenditures on the value added of the agricultural sector in Iran was investigated for the period 1971-2021. For data analysis, the researchers utilized the ARDL econometric model and EViews software. The results indicated that R&D expenditures, both in the short and long run, have a significant positive effect on the value added in the agricultural sector. The estimated elasticity coefficient for R&D expenditures in the short run was 0.45 and in the long run was 0.35, indicating that with a 1 percent increase in research and development expenditures, the value added in the agricultural sector would increase by 0.45 percent in the short run and by 0.35 percent in the long run. Moreover, variables such as capital stock, number of employees in the agricultural sector, and working days also had a significant and positive effect on the value added in the agricultural sector.
Climate change nowadays became the most problematic matter including in agricultural industries. Agriculture area productivity affected national food security and a county’s economic development. As an agricultural county, Indonesia must be ready to adapt and prepare for the worst impact of climate change. This paper aims to explore the impact of climate change on financial sustainability in agricultural industries. This research uses a systematic literature review method related to financial sustainability, climate change impact, and agriculture industries. The result shows that financial sustainability in agricultural industries must be impacted by climate change. The impact of climate change on agriculture industries is associated with reducing profitability, destroying capital, portfolio reallocation, and financial instability. Climate change caused environmental uncertainty that affects agricultural productivity. To reduce the impact of climate change on financial sustainability in agricultural industries, there must be a design of mitigation must be prepared and realized so agricultural industries are more prepared and ready to face climate change impact.
Kadyrzhan Makangali, Tamara Tultabayeva, Galia Zamaratskaia
et al.
This study investigated <i>Portulaca oleracea</i> (purslane) as a potential antioxidant supplement in cooked sausages, focusing on its effects on lipid oxidation, fatty acid composition, and antioxidant activity. The fatty acid profile of the sausages enriched with 1.2% purslane powder revealed a 1.3-fold increase in alpha-linolenic acid (ALA), an essential omega-3 fatty acid. Improved oxidative stability during refrigerated storage was observed, with peroxide values of 10.9 meq/kg in the sausages with purslane by day 10 compared with 12.5 meq/kg in the control sausages. The thiobarbituric acid (TBA) values, reflecting lipid peroxidation, were also significantly lower in the sausages with purslane. The antioxidant capacity of the sausages containing purslane was significantly enhanced, demonstrating a ferric-reducing antioxidant power (FRAP) of 13.5 mg GAE/g, whereas the control sausages showed undetectable FRAP levels. Additionally, the DPPH radical-scavenging activity in the sausages with purslane was 21.70% compared with 13.73% in the control. These findings suggest that purslane improves the nutritional profile of meat products by increasing beneficial fatty acids while providing substantial protection against oxidative spoilage. Purslane offers a promising natural alternative to synthetic antioxidants, enhancing the shelf life and quality of processed meats.
Ihor Rumyk, Tetiana Galetska, Oleksandr Klymchuk
et al.
The functioning of business structures in the agricultural sector has recently become significantly more complicated. In today's conditions, it is becoming more and more difficult to develop security strategies for agribusiness enterprises, because factors that have arisen relatively recently, especially of an external nature, have a significant negative impact on the financial activities of enterprises. In order to make optimal financial decisions, the entire toolkit, developed and tested by many years of business experience in the agricultural sector both abroad and in the middle of our country, should be used. One of these methods is economic descriptive cognitive modeling, which allows to analyze external and internal factors influencing the activities of enterprises, evaluate the strength of their interaction, and graphically display cause-and-effect relationships in a dynamic system. The method of cognitive modeling was studied in order to formalize the influence of exogenous and endogenous processes on the financial activity of agribusiness enterprises. As a result of the research, the components of the development of enterprises in the agricultural sector were analyzed from the standpoint of ensuring the efficiency of their financial activities using cognitive modeling. A matrix of causality and a cognitive map of the influence of a number of factors on the target component "financial activity of the agribusiness enterprises" were built. Impulse modeling of the influence of given concepts was carried out. The results of the conducted cognitive modeling of the influence of factors can be used to develop a safe strategy for the sustainable development of enterprises of other industries in the conditions of dynamic changes. The application of the cognitive approach made it possible to foresee various processes of development of situations in this system that may arise in it under the expected influence of various factors, as well as the influence of regulatory and control systems.
Ashwin Natraj Arun, Byunghyun Lee, Fabio A. Castiblanco
et al.
One of the most intriguing 6G vertical markets is precision agriculture, where communications, sensing, control, and robotics technologies are used to improve agricultural outputs and decrease environmental impact. Ambient IoT (A-IoT), which uses a network of devices that harvest ambient energy to enable communications, is expected to play an important role in agricultural use cases due to its low costs, simplicity, and battery-free (or battery-assisted) operation. In this paper, we review the use cases of precision agriculture and discuss the challenges. We discuss how A-IoT can be used for precision agriculture and compare it with other ambient energy source technologies. We also discuss research directions related to both A-IoT and precision agriculture.
The integration of artificial intelligence into agricultural practices, specifically through Consultation on Intelligent Agricultural Machinery Management (CIAMM), has the potential to revolutionize efficiency and sustainability in farming. This paper introduces a novel approach that leverages large language models (LLMs), particularly GPT-4, combined with multi-round prompt engineering to enhance decision-making processes in agricultural machinery management. We systematically developed and refined prompts to guide the LLMs in generating precise and contextually relevant outputs. Our approach was evaluated using a manually curated dataset from various online sources, and performance was assessed with accuracy and GPT-4 Scores. Comparative experiments were conducted using LLama-2-70B, ChatGPT, and GPT-4 models, alongside baseline and state-of-the-art methods such as Chain of Thought (CoT) and Thought of Thought (ThoT). The results demonstrate that our method significantly outperforms these approaches, achieving higher accuracy and relevance in generated responses. This paper highlights the potential of advanced prompt engineering techniques in improving the robustness and applicability of AI in agricultural contexts.
Mikolaj Cieslak, Umabharathi Govindarajan, Alejandro Garcia
et al.
We present a specialized procedural model for generating synthetic agricultural scenes, focusing on soybean crops, along with various weeds. This model is capable of simulating distinct growth stages of these plants, diverse soil conditions, and randomized field arrangements under varying lighting conditions. The integration of real-world textures and environmental factors into the procedural generation process enhances the photorealism and applicability of the synthetic data. Our dataset includes 12,000 images with semantic labels, offering a comprehensive resource for computer vision tasks in precision agriculture, such as semantic segmentation for autonomous weed control. We validate our model's effectiveness by comparing the synthetic data against real agricultural images, demonstrating its potential to significantly augment training data for machine learning models in agriculture. This approach not only provides a cost-effective solution for generating high-quality, diverse data but also addresses specific needs in agricultural vision tasks that are not fully covered by general-purpose models.
Margaret Capetz, Swati Sharma, Rafael Padilha
et al.
Mitigating climate change requires transforming agriculture to minimize environ mental impact and build climate resilience. Regenerative agricultural practices enhance soil organic carbon (SOC) levels, thus improving soil health and sequestering carbon. A challenge to increasing regenerative agriculture practices is cheaply measuring SOC over time and understanding how SOC is affected by regenerative agricultural practices and other environmental factors and farm management practices. To address this challenge, we introduce an AI-driven Soil Organic Carbon Copilot that automates the ingestion of complex multi-resolution, multi-modal data to provide large-scale insights into soil health and regenerative practices. Our data includes extreme weather event data (e.g., drought and wildfire incidents), farm management data (e.g., cropland information and tillage predictions), and SOC predictions. We find that integrating public data and specialized models enables large-scale, localized analysis for sustainable agriculture. In comparisons of agricultural practices across California counties, we find evidence that diverse agricultural activity may mitigate the negative effects of tillage; and that while extreme weather conditions heavily affect SOC, composting may mitigate SOC loss. Finally, implementing role-specific personas empowers agronomists, farm consultants, policymakers, and other stakeholders to implement evidence-based strategies that promote sustainable agriculture and build climate resilience.
Muhammad Awais, Ali Husain Salem Abdulla Alharthi, Amandeep Kumar
et al.
Significant progress has been made in advancing large multimodal conversational models (LMMs), capitalizing on vast repositories of image-text data available online. Despite this progress, these models often encounter substantial domain gaps, hindering their ability to engage in complex conversations across new domains. Recent efforts have aimed to mitigate this issue, albeit relying on domain-specific image-text data to curate instruction-tuning data. However, many domains, such as agriculture, lack such vision-language data. In this work, we propose an approach to construct instruction-tuning data that harnesses vision-only data for the agriculture domain. We utilize diverse agricultural datasets spanning multiple domains, curate class-specific information, and employ large language models (LLMs) to construct an expert-tuning set, resulting in a 70k expert-tuning dataset called AgroInstruct. Subsequently, we expert-tuned and created AgroGPT, an efficient LMM that can hold complex agriculture-related conversations and provide useful insights. We also develop AgroEvals for evaluation and compare {AgroGPT's} performance with large open and closed-source models. {AgroGPT} excels at identifying fine-grained agricultural concepts, can act as an agriculture expert, and provides helpful information for multimodal agriculture questions. The code, datasets, and models are available at https://github.com/awaisrauf/agroGPT.