Shun Hattori, Hikaru Sasaki, Takumi Hachimine
et al.
Vision-based imitation learning has shown promise for robotic manipulation; however, its generalization remains limited in practical agricultural tasks. This limitation stems from scarce demonstration data and substantial visual domain gaps caused by i) crop-specific appearance diversity and ii) background variations. To address this limitation, we propose Dual-Region Augmentation for Imitation Learning (DRAIL), a region-aware augmentation framework designed for generalizable vision-based imitation learning in agricultural manipulation. DRAIL explicitly separates visual observations into task-relevant and task-irrelevant regions. The task-relevant region is augmented in a domain-knowledge-driven manner to preserve essential visual characteristics, while the task-irrelevant region is aggressively randomized to suppress spurious background correlations. By jointly handling both sources of visual variation, DRAIL promotes learning policies that rely on task-essential features rather than incidental visual cues. We evaluate DRAIL on diffusion policy-based visuomotor controllers through robot experiments on artificial vegetable harvesting and real lettuce defective leaf picking preparation tasks. The results show consistent improvements in success rates under unseen visual conditions compared to baseline methods. Further attention analysis and representation generalization metrics indicate that the learned policies rely more on task-essential visual features, resulting in enhanced robustness and generalization.
Mohammed Brahimi, Karim Laabassi, Mohamed Seghir Hadj Ameur
et al.
Machine learning models in agricultural vision often achieve high accuracy on curated datasets but fail to generalize under real field conditions due to distribution shifts between training and deployment environments. Moreover, most machine learning competitions focus primarily on model design while treating datasets as fixed resources, leaving the role of data collection practices in model generalization largely unexplored. We introduce the AgrI Challenge, a data-centric competition framework in which multiple teams independently collect field datasets, producing a heterogeneous multi-source benchmark that reflects realistic variability in acquisition conditions. To systematically evaluate cross-domain generalization across independently collected datasets, we propose Cross-Team Validation (CTV), an evaluation paradigm that treats each team's dataset as a distinct domain. CTV includes two complementary protocols: Train-on-One-Team-Only (TOTO), which measures single-source generalization, and Leave-One-Team-Out (LOTO), which evaluates collaborative multi-source training. Experiments reveal substantial generalization gaps under single-source training: models achieve near-perfect validation accuracy yet exhibit validation-test gaps of up to 16.20% (DenseNet121) and 11.37% (Swin Transformer) when evaluated on datasets collected by other teams. In contrast, collaborative multi-source training dramatically improves robustness, reducing the gap to 2.82% and 1.78%, respectively. The challenge also produced a publicly available dataset of 50,673 field images of six tree species collected by twelve independent teams, providing a diverse benchmark for studying domain shift and data-centric learning in agricultural vision.
As the volume of unstructured text continues to grow across domains, there is an urgent need for scalable methods that enable interpretable organization, summarization, and retrieval of information. This work presents a unified framework for interpretable topic modeling, zero-shot topic labeling, and topic-guided semantic retrieval over large agricultural text corpora. Leveraging BERTopic, we extract semantically coherent topics. Each topic is converted into a structured prompt, enabling a language model to generate meaningful topic labels and summaries in a zero-shot manner. Querying and document exploration are supported via dense embeddings and vector search, while a dedicated evaluation module assesses topical coherence and bias. This framework supports scalable and interpretable information access in specialized domains where labeled data is limited.
Promoting the integrated development of rural industries represents a crucial pathway for accelerating the modernization of the entire industrial chain and consolidating and enhancing agricultural resilience. This study constructs provincial-level panel data for China spanning 2012–2022 and employs a multidimensional, visualization-based, and spatial research paradigm to comprehensively examine the impact effects and mechanisms through which rural industrial integration empowers agricultural resilience in China. The findings reveal that, first, rural industrial integration can effectively enhance agricultural resilience levels, with stable economic returns and diversified functional development serving as key pathways for improving agricultural resilience. Second, the eastern and western regions have unleashed agricultural resilience potential through superior policy and environmental advantages, while the relatively homogeneous grain structure and path dependence in the central and northeastern regions have prevented agricultural industrial integration from effectively achieving expected outcomes. Third, structural rationalization has improved resource allocation efficiency, but industrial structure advancement and industrial structure sophistication have actually diminished the enhancement effects on agricultural resilience. Fourth, urbanization has led to adverse resource competition, reducing the positive impact of rural industrial integration on agricultural resilience levels. Fifth, the enhancement of agricultural resilience through rural industrial integration demonstrates geographically-distant resource spillover effects to neighboring regions. This research provides an important theoretical framework and practical paradigm for global agricultural transformation, offering particularly significant strategic guidance for developing countries in building resilient agricultural systems, addressing climate change impacts, and ensuring food security.
Accurate mapping of individual trees is an important component for precision agriculture in orchards, as it allows autonomous robots to perform tasks like targeted operations or individual tree monitoring. However, creating these maps is challenging because GPS signals are often unreliable under dense tree canopies. Furthermore, standard Simultaneous Localization and Mapping (SLAM) approaches struggle in orchards because the repetitive appearance of trees can confuse the system, leading to mapping errors. To address this, we introduce Tree-SLAM, a semantic SLAM approach tailored for creating maps of individual trees in orchards. Utilizing RGB-D images, our method detects tree trunks with an instance segmentation model, estimates their location and re-identifies them using a cascade-graph-based data association algorithm. These re-identified trunks serve as landmarks in a factor graph framework that integrates noisy GPS signals, odometry, and trunk observations. The system produces maps of individual trees with a geo-localization error as low as 18 cm, which is less than 20\% of the planting distance. The proposed method was validated on diverse datasets from apple and pear orchards across different seasons, demonstrating high mapping accuracy and robustness in scenarios with unreliable GPS signals.
Kashif Sattar, Muhammad Arslan, Saqib Majeed
et al.
Wireless Sensor Networks have risen as a highly promising technology suitable for precision agriculture implementations, enabling efficient monitoring and control of agricultural processes. In precision agriculture, accurate and synchronized data collection is crucial for effective analysis and decision making. Using principles of information theory, we can define conditions and parameters that influence the efficient transmission and processing of information. Existing technologies have limitations in maintaining consistent time references, handling node failures, and unreliable communication links, leading to inaccurate data readings. Reliable data storage is demanding now-a-days for storing data on local monitoring station as well as in online live server. Sometime internet is not working properly due to congestion and there is frequent packet loss. Current solutions often synchronize records based on database timestamps, leading to record duplication and waste storage. Both databases synchronize each other after internet restoration. By providing synchronization among nodes and data, accuracy and storage will be saved in IoT based WSNs for precision agriculture applications. A prototype Node-MCU internal memory is used as a resource for achieving data synchronization. This proposed work generates record ID from Node MCU EEPROM which helps in records synchronization if there is any packet loss at the local server or at the online server to maintain synchronization accuracy despite unreliable communication links. Experiment shows that for a particular duration Node MCU generated 2364 packets and packet loss at local server was 08 and at online server was 174 packets. Results shows that after synchronization 99.87% packets were synchronized. Using previous technique of timestamp, the redundancy was 70% which reduced to 0% using our proposed technique.
Ammonia (NH3) emissions significantly contribute to atmospheric pollution, yet discrepancies exist between bottom-up inventories and satellite-constrained top-down estimates, with the latter typically one-third higher. This study quantifies how assumptions about NH3 vertical distribution in satellite retrievals contribute to this gap. By implementing spatially and temporally resolved vertical profiles from the Community Multiscale Air Quality model to replace steep gradients in Infrared Atmospheric Sounding Interferometer (IASI) retrievals, we reduced satellite-model column discrepancies from 71% to 18%. We subsequently constrained NH3 emissions across China using a hybrid inversion framework combining iterative mass balance and four-dimensional variational methods. Our posterior emissions showed agreement with the a priori inventory (7.9% lower), suggesting that discrepancies between inventory approaches were amplified by overestimation of near-surface NH3 in baseline satellite retrievals, potentially causing a 43% overestimation of growing season emissions. Evaluation against ground-based measurements confirmed improved model performance, with normalized root-mean-square error reductions of 1-27% across six months. These findings demonstrate that accurate representation of vertical profiles in satellite retrievals is critical for robust NH3 emission estimates and can reconcile the long-standing discrepancy between bottom-up and top-down approaches. Our hybrid inversion methodology, leveraging profile-corrected satellite data, reveals that China's NH3 emissions exhibit greater spatial concentration than previously recognized, reflecting agricultural intensification. This advancement enables timely and accurate characterization of rapidly changing agricultural emission patterns, critical for implementing effective nitrogen pollution control measures.
Marloes P. van Loon, Seyyedmajid Alimagham, Isaac K. Abuley
et al.
Europe is an important potato producer, showing a strong decline in areas and increases in yield over the past decades, but with large regional differences. This study aims to characterise current European potato production by analysing yields, revealing yield gaps (Yg), and assessing key factors that explain actual (Ya) and potential yields (Yw, for rainfed systems; Yp, for irrigated systems). We selected 13 key potato producing countries, jointly accounting for 90% of the European potato area. Local data were used to simulate Yw and Yp, while Ya was retrieved from sub-national statistics. Then, we analysed main factors affecting yields using boundary line analysis on nitrogen input and crop water availability. Results showed that European potato production on current acreage can increase by 55% when yields would increase to 80% of their potential. The largest potential production gains featured in eastern Europe (59% Yg, 59% of potato area), thereafter western Europe (32% Yg, 25% of potato area), and smallest gains in northern and southern Europe (43% and 45% Yg, with relatively small acreages of 9% and 6%, respectively). Our analysis revealed that nitrogen input was a limiting factor in eastern Europe, while we found substantial overuse in some western European countries. Under rainfed conditions, water was the main limiting factor in relatively few potato cultivation areas. In irrigated areas, e.g. in southern Europe, irrigation water requirements to approach Yp are large, which becomes increasingly challenging. Insights from this study can be used to guide future development and innovation in potato cultivation across Europe.
Line Vinther Hansen, Azeem Tariq, Lars Stoumann Jensen
et al.
Choosing the appropriate timing of fertilisation is one of the primary managerial strategies to avoid emissions of nitrous oxide. To minimise odour nuisance and ammonia volatilisation, farmers are advised to apply pig slurry (PS) before light rainfall events. However, application of manure at the time of heavy rain events can pose a high risk of nitrous oxide (N2O) emissions, and analysing weather forecasts to avoid this could be important to mitigate emissions. A controlled field experiment was conducted to assess the effect of rainfall around the time of PS fertilisation on soil N2O emissions. The main findings were: 1) Rainfall treatments showed slight peaks in N2O emissions up to 3.9 mg N2O-N ha−1 day−1 after a rain event, and cumulative N2O emissions were numerically higher compared to the treatment without rainfall, suggesting a higher risk when manure application coincides with rainfall. 2) However, the short duration of elevated soil water content with more than 80 % water-filled pore space (WFPS) lasting only a few days after a rain event, may not have been sufficient to create anoxic conditions necessary for substantial N2O emissions from PS. This study's simulated rain approach can be used for future manipulative studies of field sites to further investigate mitigation options related to rainfall to inform best practices for fertiliser application in relation to weather forecasts.
Abstract Short food supply chains (SFSCs) are increasingly regarded as promising alternatives to industrialized food distribution systems. They aim to create geographical, logistical, and social proximity between food producers and consumers. Despite extensive research on SFSCs in recent years, our understanding of how they can best be promoted is hindered by the lack of a unified approach in extant studies of consumer perspectives on SFSCs. The aim of this systematic literature review is thus to provide a comprehensive overview of the key determinants shaping consumer behavior around SFSCs. Following PRISMA guidelines, 30 peer-reviewed articles based on empirical research of SFSC consumers in the European Union were selected for analysis, applying Alphabet Theory as a theoretical framework. The findings emphasize the influence on consumer behavior of contextual factors, information availability, and trust, as well as the relevance of other commonly studied factors such as attitudes and sociodemographics. The review further highlights the importance of understanding how these factors interact to shape consumers’ perceptions of the costs and benefits of SFSCs. Based on these findings, the study gives recommendations to address the challenges identified and suggests directions for future research.
Nutrition. Foods and food supply, Agricultural industries
Spatial variability analysis is a pivotal component of precision agriculture, particularly in the assessment of soil health. This research focuses on a rapeseed field located in western Iran, underscoring the critical role of spatial variability in evaluating soil quality and agricultural productivity. A comprehensive sampling strategy involved collecting undisturbed soil samples from 245 strategically chosen points within a 50 m x 50 m grid, encompassing both surface and subsurface layers. Key indicators for assessing the soil's physical quality included available water content, saturated hydraulic conductivity (Kfs), bulk density (BD), noncapillary porosity and organic carbon levels. The physical rating index (PRI), a key metric for measuring soil productivity, was derived from the integrated rating values of these parameters. After the growing season, manual harvesting of rapeseed allowed for the quantification of biological yield. To create pedotransfer functions (PTF) for estimating bulk density (BD) and saturated hydraulic conductivity (Kfs), multiple regression analysis was utilized. Findings indicated the existence of a subsurface hardpan attributable to soil compaction, which impedes root development and negatively impacts soil productivity. By evaluating critical soil parameters, specifically, BD (>1.6 Mg m-3) and Kfs (<0.5 cm hr-1), the compacted zones were identified within the field. Geostatistical analysis, particularly kriging interpolation, facilitated the creation of spatial maps for BD, Kfs, and PRI, equipping farmers with targeted insights for deep tillage interventions aimed at enhancing soil health. Ultimately, this study elucidates how spatial variability analysis can inform agricultural practices, optimize crop yields, and reduce input costs.
Salima Yousfi, Mohammad Shahid, Sumitha Thushar
et al.
Subtropical arid regions face challenges such as high temperatures, poor soil fertility, and saline soils and water. Quinoa (Chenopodium quinoa Willd.) is well-suited for these areas, particularly in the Middle East and North Africa, where it is cultivated with saline irrigation. This study evaluated the seed yield and quality of eleven quinoa genotypes under two salinity levels (1 and 15 dSm−1) in sandy soils at the International Center for Biosaline Agriculture (Dubai, UAE). Key traits measured included seed yield (SY), biomass, plant height (PH), days to flowering (DF) and maturity (DM), thousand seed weight (TSW), and carbon (δ13C) and nitrogen (δ15N) isotope composition, alongside nitrogen, mineral, and essential amino acid content. Salinity reduced PH and biomass by 20 %, TSW by 6 %, and SY by 30 %, although the ICBA-Q5 variety showed positive effects on SY and biomass under saline conditions. Salinity increased δ13C and δ15N and had a minor impact on minerals, with moderate increases in zinc and sulfur. Amino acids showed slight reductions in isoleucine, leucine, and threonine. Genotypic effects were more significant than salinity, with Ames-13757 performing best for SY and biomass under both control and saline conditions, and NSL-106399 displaying the highest amino acid content under salinity. δ13C and DF negatively correlated with SY, while δ15N was linked negatively to some minerals and amino acids. The study found a low trade-off between seed yield and quality under salinity. Quinoa is highly adaptable to saline irrigation in the UAE, and genotype selection is key for optimizing yield and quality.
Mohammad Zia Ur Rehman, Devraj Raghuvanshi, Nagendra Kumar
Delivering prompt information and guidance to farmers is critical in agricultural decision-making. Farmers helpline centres are heavily reliant on the expertise and availability of call centre agents, leading to inconsistent quality and delayed responses. To this end, this article presents Kisan Query Response System (KisanQRS), a Deep Learning-based robust query-response framework for the agriculture sector. KisanQRS integrates semantic and lexical similarities of farmers queries and employs a rapid threshold-based clustering method. The clustering algorithm is based on a linear search technique to iterate through all queries and organize them into clusters according to their similarity. For query mapping, LSTM is found to be the optimal method. Our proposed answer retrieval method clusters candidate answers for a crop, ranks these answer clusters based on the number of answers in a cluster, and selects the leader of each cluster. The dataset used in our analysis consists of a subset of 34 million call logs from the Kisan Call Centre (KCC), operated under the Government of India. We evaluated the performance of the query mapping module on the data of five major states of India with 3,00,000 samples and the quantifiable outcomes demonstrate that KisanQRS significantly outperforms traditional techniques by achieving 96.58% top F1-score for a state. The answer retrieval module is evaluated on 10,000 samples and it achieves a competitive NDCG score of 96.20%. KisanQRS is useful in enabling farmers to make informed decisions about their farming practices by providing quick and pertinent responses to their queries.
Yash Zambre, Ekdev Rajkitkul, Akshatha Mohan
et al.
Object detection plays a crucial role in the field of computer vision by autonomously locating and identifying objects of interest. The You Only Look Once (YOLO) model is an effective single-shot detector. However, YOLO faces challenges in cluttered or partially occluded scenes and can struggle with small, low-contrast objects. We propose a new method that integrates spatial transformer networks (STNs) into YOLO to improve performance. The proposed STN-YOLO aims to enhance the model's effectiveness by focusing on important areas of the image and improving the spatial invariance of the model before the detection process. Our proposed method improved object detection performance both qualitatively and quantitatively. We explore the impact of different localization networks within the STN module as well as the robustness of the model across different spatial transformations. We apply the STN-YOLO on benchmark datasets for Agricultural object detection as well as a new dataset from a state-of-the-art plant phenotyping greenhouse facility. Our code and dataset are publicly available.
Nathan P. Lawrence, Seshu Kumar Damarla, Jong Woo Kim
et al.
With the rise of deep learning, there has been renewed interest within the process industries to utilize data on large-scale nonlinear sensing and control problems. We identify key statistical and machine learning techniques that have seen practical success in the process industries. To do so, we start with hybrid modeling to provide a methodological framework underlying core application areas: soft sensing, process optimization, and control. Soft sensing contains a wealth of industrial applications of statistical and machine learning methods. We quantitatively identify research trends, allowing insight into the most successful techniques in practice. We consider two distinct flavors for data-driven optimization and control: hybrid modeling in conjunction with mathematical programming techniques and reinforcement learning. Throughout these application areas, we discuss their respective industrial requirements and challenges. A common challenge is the interpretability and efficiency of purely data-driven methods. This suggests a need to carefully balance deep learning techniques with domain knowledge. As a result, we highlight ways prior knowledge may be integrated into industrial machine learning applications. The treatment of methods, problems, and applications presented here is poised to inform and inspire practitioners and researchers to develop impactful data-driven sensing, optimization, and control solutions in the process industries.
Pierre La Rocca, Gaël Guennebaud, Aurélie Bugeau
et al.
Digitalization appears as a lever to enhance agriculture sustainability. However, existing works on digital agriculture's own sustainability remain scarce, disregarding the environmental effects of deploying digital devices on a large-scale. We propose a bottom-up method to estimate the carbon footprint of digital agriculture scenarios considering deployment of devices over a diversity of farm sizes. It is applied to two use-cases and demonstrates that digital agriculture encompasses a diversity of devices with heterogeneous carbon footprints and that more complex devices yield higher footprints not always compensated by better performances or scaling gains. By emphasizing the necessity of considering the multiplicity of devices, and the territorial distribution of farm sizes when modelling digital agriculture deployments, this study highlights the need for further exploration of the first-order effects of digital technologies in agriculture.
The increasing impact of global climate change has resulted in adversity stresses, like salt and drought, gradually becoming the main factors that limit crop growth. Hemp, which contains numerous medicinal active components and multiple bioactive functions, is widely used in the agricultural, industrial, and medical fields, hence promoting the rapid development of related industries. Arbuscular mycorrhizal fungi (AMF) can establish a symbiotic relationship with 80% of vascular plants. This symbiosis promotes host plant growth, regulates plant physiology and biochemistry, facilitates secondary metabolite synthesis, and enhances resistance to abiotic stresses. However, the effects of salt stress, drought stress, and AMF interaction in hemp are not well understood. In this study, to investigate this, we performed a study where we cultured hemp that was either inoculated or uninoculated with <i>Funneliformis mosseae</i> and determined changes in effective colonization rate, growth, soluble substances, photosynthesis, fluorescence, ions, and secondary metabolites by cultivating hemp under different concentrations of NaCl (0 mM, 100 mM, and 200 mM) and different soil moisture content (45%, 25%, and 15%). The results showed that salt, drought stress, or salt–drought interaction stress all inhibited colonization rate after stress, plant growth, mainly due to ion toxicity and oxidative damage. Inoculation with <i>F. mosseae</i> effectively alleviated plant growth inhibition under 100 mM NaCl salt stress, drought stress, and salt–drought interaction stress conditions. It also improved osmoregulation, photosynthetic properties, fluorescence properties, and ion homeostasis, and promoted the accumulation of secondary metabolites. However, under 200 mM NaCl salt stress conditions, inoculation with <i>F. mosseae</i> negatively affected plant physiology, biochemistry, and secondary metabolite synthesis, although it did alleviate growth inhibition. The results demonstrate that there are different effects of salt–drought interaction stress versus single stress (salt or drought stress) on plant growth physiology. In addition, we provide new insights about the positive effects of AMF on host plants under such stress conditions and the effects of AMF on plants under high salt stress.
Keartisak Sriprateep, Surajet Khonjun, Paulina Golinska-Dawson
et al.
The classification of certain agricultural species poses a formidable challenge due to their inherent resemblance and the absence of dependable visual discriminators. The accurate identification of these plants holds substantial importance in industries such as cosmetics, pharmaceuticals, and herbal medicine, where the optimization of essential compound yields and product quality is paramount. In response to this challenge, we have devised an automated classification system based on deep learning principles, designed to achieve precision and efficiency in species classification. Our approach leverages a diverse dataset encompassing various cultivars and employs the Parallel Artificial Multiple Intelligence System–Ensemble Deep Learning model (P-AMIS-E). This model integrates ensemble image segmentation techniques, including U-Net and Mask-R-CNN, alongside image augmentation and convolutional neural network (CNN) architectures such as SqueezeNet, ShuffleNetv2 1.0x, MobileNetV3, and InceptionV1. The culmination of these elements results in the P-AMIS-E model, enhanced by an Artificial Multiple Intelligence System (AMIS) for decision fusion, ultimately achieving an impressive accuracy rate of 98.41%. This accuracy notably surpasses the performance of existing methods, such as ResNet-101 and Xception, which attain 93.74% accuracy on the testing dataset. Moreover, when applied to an unseen dataset, the P-AMIS-E model demonstrates a substantial advantage, yielding accuracy rates ranging from 4.45% to 31.16% higher than those of the compared methods. It is worth highlighting that our heterogeneous ensemble approach consistently outperforms both single large models and homogeneous ensemble methods, achieving an average improvement of 13.45%. This paper provides a case study focused on the Centella Asiatica Urban (CAU) cultivar to exemplify the practical application of our approach. By integrating image segmentation, augmentation, and decision fusion, we have significantly enhanced accuracy and efficiency. This research holds theoretical implications for the advancement of deep learning techniques in image classification tasks while also offering practical benefits for industries reliant on precise species identification.
Headland maneuvering is a crucial aspect of unmanned field operations for autonomous agricultural vehicles (AAVs). While motion planning for headland turning in open fields has been extensively studied and integrated into commercial auto-guidance systems, the existing methods primarily address scenarios with ample headland space and thus may not work in more constrained headland geometries. Commercial orchards often contain narrow and irregularly shaped headlands, which may include static obstacles,rendering the task of planning a smooth and collision-free turning trajectory difficult. To address this challenge, we propose an optimization-based motion planning algorithm for headland turning under geometrical constraints imposed by field geometry and obstacles.