Blockchain has been increasingly explored as a tool to enhance trust, transparency, and efficiency in agricultural data management. Existing review papers in this field primarily catalog blockchain studies, platform designs, and emerging applications in agriculture, yet they rarely interrogate the deeper challenges of data sharing that limit blockchain’s practical adoption. This study goes beyond descriptive review to analyze the structural gaps that arise when agricultural data must circulate across heterogeneous actors, workflows, and jurisdictions. An interpretive cross-sectoral comparison of blockchain implementations in agriculture, healthcare, and finance was conducted to indentify which data-sharing functions remain absent in agriculture. Three persistent shortcomings in agriculture: the absence of enforceable usage rights, incomplete verification of data transformations, and lack of cross-system interoperability. The absence of these functions, which were more clearly implemented in the reviewed healthcare and finance studies, implies that agricultural data can lose provenance, permissions, and meaning once reused beyond their original platforms. Building on these findings, the study outlines a conceptual dual smart contract design that brings these missing functions together as an architectural direction for agricultural data sharing.
Weeds significantly reduce crop productivity and profitability, particularly in sugar beet production where losses exceed 1.25 billion annually in the United States. Limitations of herbicide-based control and drawbacks of conventional tillage highlight the need for efficient, sustainable alternatives. To develop and evaluate an autonomous site-specific mechanical weeding (SSMW) system with adaptive control for precise inter-row weed removal under field conditions. An SSMW system was designed by integrating an inter-row tillage mechanism with a robotic platform using RTK-GPS for navigation and site-specific actuation. Site-specific weeding was performed in the field test and a strain-based feedback control system was implemented in computer simulation to enable ASIC tillage under uneven terrain. Simulation studies based on field data were conducted to validate control performance, followed by field experiments to assess system effectiveness. The SSMW system’s efficacy was found to be 96.5
In-season optical sensor-based variable-rate nitrogen application (VRA) is designed to optimize nitrogen (N) management in maize (Zea mays L.) by aligning N inputs with real-time crop demand, yet reports on its performance are inconsistent. This study quantified the effects of VRA relative to fixed-rate management on N inputs, yield, N use efficiency, and profit; assessed VRA’s economic response to market volatility; identified key moderators; tested hierarchical interactions among moderators; and characterized economic and N-efficiency trade-offs. A systematic review through February 2026 yielded 25 studies and 235 paired comparisons. The log response ratio was calculated for total N rate, grain yield, partial factor productivity of N (PFPN), and partial profit. Weighted linear mixed-effects models estimated overall effects, single-moderator analyses identified drivers, and conditional inference trees identified hierarchical interactions. VRA reduced total N rate by 18
Purpose Accurate measurement of corn stalk diameter is important for assessing plant robustness, lodging resistance, and harvest performance, but automated measurement across changing crop conditions remains challenging. This study evaluated whether a ground-based stereo-vision system could reliably estimate stalk diameter throughout the growing season, from mid-summer through senescence. Methods A stereo-vision pipeline using dual AR0234 global-shutter cameras was deployed on an autonomous ground robot. YOLOv8 provided stalk localization and pose correction, BoT-SORT enabled multi-frame tracking, and U-Net segmentation with an edge-mask strategy extracted stalk boundaries under partial occlusion. Repeated per-frame width estimates were filtered and converted to physical dimensions using disparity-based calibration. Vision estimates were compared with perpendicular caliper measurements. Results Under mid-summer conditions with minimal leaf interference, the system achieved a mean absolute error (MAE) of 1.1–1.5 mm and r² of 0.90. During late-season senescence, leaf-sheath expansion and occlusion caused systematic diameter overestimation and reduced accuracy. Applying a seasonally derived offset of approximately 3.8 mm reduced MAE to approximately 1.3 mm, although correlation remained modest (r² ≈ 0.41). Independent human measurements also exhibited variability, with inter-rater r² ≈ 0.77 and mean disagreement of approximately 1.0 mm. Conclusion Stereo vision can provide accurate, non-contact corn stalk diameter measurements under favorable canopy conditions and remains viable across the crop lifecycle. However, robust late-season phenotyping will require improved modeling of leaf sheaths and occlusions together with more reliable ground-truth measurement procedures.
Evaluate how soil and canopy sensing can map within-block variability in tart cherry orchards and identify indicators robust enough for repeatable management decisions. Soil apparent electrical conductivity (ECa) was mapped in spring 2022 across four commercial tart cherry blocks (8.5–10.5 ha; approximately 3,500 trees per block), followed by canopy sensing in 2023–2024. Canopy structure was measured using unmanned aerial vehicle (UAV) photogrammetry and mobile terrestrial laser scanning (MTLS) using light detection and ranging (LiDAR), and canopy density using mobile ceptometry. Spatial layers were aligned to per-tree grid cells. An August 2025 campaign compared UAV- and LiDAR-derived tree height with ground-truthed height. Soil-to-canopy relationships were weak to moderate but consistent within blocks (r = 0.10–0.40), with strength and direction varying by site conditions. Canopy density was more strongly associated with UAV-derived volume than height. UAV-derived 90th-percentile height best predicted ground-truthed height (R² = 0.89; RMSE = 0.34 m), whereas LiDAR showed a weaker relationship and greater error (R² = 0.70; RMSE = 0.52 m). Cross-sensor agreement was moderate to strong (r = 0.41–0.65). Per-tree rankings were stable between years for UAV height and volume. Whole-block sensing revealed persistent spatial patterns that could support management-zone delineation. UAV photogrammetry provided accurate canopy metrics, MTLS offered measurements suited to routine orchard operations, ceptometry added seasonal canopy-density information, and ECa provided soil context. Occasional ECa mapping combined with strategically timed UAV surveys and other sensors as needed could reduce redundant sensing while supporting fertilizer evaluation, pruning, and labor allocation. Multi-sensor integration: Combined soil apparent electrical conductivity, UAV photogrammetry, LiDAR, and mobile ceptometry to map within-orchard variability at commercial scale. Cross-sensor evaluation: Quantified relationships among soil, canopy structure, and canopy density; UAV canopy height (P90) most accurately predicted ground truth (R² = 0.89). Management relevance: Identified repeatable spatial patterns that could support zone-based, variable-rate, and labor-efficient management strategies in tart cherry orchards. This study demonstrates how integrating mapped soil apparent electrical conductivity, UAV photogrammetry, LiDAR, and mobile ceptometry can quantify within-orchard variability at commercial scale. Linking soil and canopy information provides a practical framework for identifying persistent spatial patterns that could support management-zone delineation and guide variable-rate or labor-efficient orchard practices.
Powdery mildew (PM), caused by Erysiphe betae, is the most common sugar beet leaf disease, leading to significant yield losses. Traditional disease monitoring via visual scoring is labour-intensive and prone to bias. Visible-near infrared spectroscopy, a non-invasive technique measuring plant-electromagnetic radiation interactions, can capture physiological responses to biotic stress like PM. This study evaluates spectral disease indices (SDIs) to detect and differentiate PM infection levels in sugar beet across greenhouse and field trials. Spectral data were collected using the ASD FieldSpec 4 portable spectrometer measuring across the 350–2500 nm range. Two types of SDIs were developed: normalized difference (ND) indices based on two wavelengths, obtained by exhaustive search, and ND+ indices incorporating three wavelengths, optimized using genetic algorithms. The study identified key spectral regions and wavelengths related to pigments, mesophyll structure, water, and lipids, with indices achieving high performance in differentiating no to low infection from severe infection. While early detection proved challenging due to weak spectral signals from initial symptoms, later-stage discrimination was robust across years, locations, and trials. These findings highlight the potential of SDIs for efficient, non-invasive disease monitoring in breeding programmes, enabling the development of cost-effective and transferable detection tools for precision agriculture.
Purpose Accurate detection of the maize seedling center is critical for automated management and intelligent weeding in precision agriculture. To address the challenges of low detection accuracy and poor robustness of crop center detection under complex field environments, this paper proposes a novel detection method that integrates deep learning-based instance segmentation with leaf skeleton geometric morphological analysis. Methods First, the YOLOv8n-seg model is enhanced by incorporating ODConv_3rd dynamic convolution and the WIoU loss function, enabling high-fidelity instance segmentation of maize seedling leaves while effectively suppressing weed-related misclassification in complex backgrounds. Based on the segmented leaves, leaf skeletons are extracted to fit central curves, the true plant growth center is then determined by calculating the intersection of their extension line segments. Results The enhanced YOLOv8n-seg model yielded a 3.8
L-band passive microwave radiometry enables reliable soil moisture retrieval from satellites but at spatial resolutions too coarse for precision irrigation. Portable near-ground L-band radiometers offer a high-resolution alternative. This study evaluated a Portable L-band Radiometer (PoLRa) for estimating volumetric water content (VWC), comparing manufacturer-calibrated estimates with data-driven calibration approaches and quantifying ground-truth requirements. Field experiments were conducted at four golf courses in Texas and Virginia from August 2023 to March 2025, with three fairways per site across 17 surveys. Brightness temperature was collected using a vehicle-mounted PoLRa, and ground-truth VWC was measured at 11 to 12 locations per fairway using time domain reflectometry (TDR). Model performance was evaluated across 5,000 calibration and evaluation iterations, with three samples per fairway. Site-dependent models employed analysis of covariance (ANOCOVA), and minimal calibration analyses evaluated ground-truth sample sizes (N = 2 to 8) per fairway. Manufacturer-calibrated PoLRa estimates showed weak agreement with TDR (R² = 0.40). Data-driven site-independent models failed to maintain accuracy across locations and times (RMSE > 0.089 m3 m− 3). Site-dependent models substantially improved performance (R² = 0.727 to 0.757; RMSE = 0.052 to 0.056 m3 m− 3), with the global ANOCOVA model balancing accuracy and scalability. Minimal calibration analysis showed three ground-truth samples per fairway achieved acceptable errors (RMSE < 0.06 m³ m⁻³) in 75
Long-term monitoring of crop biophysical and biochemical traits remains challenging in high-latitude regions due to short growing seasons, frequent cloud cover, and highly variable weather. In this context, unmanned aerial vehicles (UAVs) offer flexible, high-resolution observations, but their added value relative to low-cost proximal sensors and their effectiveness for radiative transfer model (RTM) inversion across diverse crop canopies remain insufficiently quantified. This study evaluated the potential of a two-band proximal spectral reflectance sensor (SRS) and a five-band multispectral UAV sensor for retrieving leaf area index (LAI), leaf chlorophyll content (LCC), and canopy chlorophyll content (CCC) using PROSAIL inversion across major crops in Northern Europe over two growing seasons (2023–2024). Two inversion approaches – look-up table (LUT) and artificial neural network (ANN) were applied to PROSAIL simulations. UAV–PROSAIL–ANN outperformed LUT-based inversion and SRS observations, achieving the highest accuracy for LAI (R2 = 0.81–0.95; RMSE = 0.27–0.77 m2/m2), followed by CCC (R2 = 0.58–0.94; RMSE < 60 μg/cm2), while LCC remained less accurately estimated (R2 = 0.26–0.78; RMSE < 16 μg/cm2). Across sensors and methods, retrieval accuracy decreased in the order of LAI, CCC, and LCC, reflecting the stronger spectral control of canopy structure compared to biochemical traits. The UAV–PROSAIL–ANN framework effectively captured spatial and temporal variability in crop traits, producing canopy-scale maps consistent with field observations. These results demonstrate the robustness and scalability of hybrid PROSAIL–ANN inversion for high-latitude crop monitoring, while highlighting current limitations in biochemical trait retrieval using multispectral data.
Weeding is a crucial agricultural practice for reducing competition between crops and invasive plants for essential resources required for growth. However, most existing deep convolutional neural networks for weed detection are developed under Closed-set supervised learning, in which all weed species encountered during deployment are assumed to belong to predefined training categories. This assumption limits their robustness in real fields, where unknown weed species frequently emerge alongside crops and known weeds. In this study, a deep nonlinear autoencoder (DNAE) prototype network based on Open-set learning is proposed for robust detection of known and unknown (KU) weeds. The DNAE module learns discriminative representations of known categories while identifying unknown weed instances that deviate from the learned known category feature space. Despite this success, the dual-classifier for known and unknown detection introduces conflicting unknown detections within crop leaf and center regions, complicating precise crop detection and localization. Therefore, a Target-Specific Unknown Suppression (TSUS) algorithm is further introduced to suppress these conflicting unknown detections. In addition, a convex-based crop center localization (C3L) method is developed to leverage the Open-set pipeline for robotic crop safety control. Experiments conducted under the proposed Open-set pipeline detect both known and unknown weed species while consistently reducing absolute Open-set error (AOSE). Under Case 2, the Open-set pipeline achieved unknown recall RU values of 17.80
Purpose Weed management in vegetable production systems is increasingly constrained by labor shortages, rising input costs, and the need to reduce herbicide use while avoiding crop injury. Precision, site-specific spraying offers a promising alternative to broadcast application; however, its effectiveness under real field conditions is often limited by unreliable weed detection, sprayer resolution, and timing inaccuracies. This study evaluates an integrated high-precision smart spraying system combining real-time weed detection, plant tracking, and micro-jet spray actuation for selective weed control in vegetable fields. Methods The system employed a YOLOv10-small model trained on a five-season crop-weed dataset (14,186 images and 103,266 annotated plant instances), coupled with a ByteTrack algorithm for spray timing. A micro-jet sprayer equipped with 12 independently controlled nozzles spaced at 1-cm intervals was mounted on a ground-based, robotic platform to target early-stage weeds, and an optimized multithreaded software architecture was implemented for system integration and real-time performance. Following an initial dataset-based crop-weed detection evaluation, the system was tested in a lettuce field plot, a 15-m crop row containing 74 lettuce plants and 174 weeds, to further evaluate plant detection and spraying performance. Blue dye-based fluid was used in the spraying testing of the system at a forward speed of 0.91 km/h. Results Video-based evaluation yielded a detection performance of 80.0
Precision agriculture (PA) is a key strategy for advancingsustainable food production under the European Green Deal and theFarm to Forkframework. While a meta-analysis of 234 studies (1988-2005) confirmed PAprofitability in 68
Soil heavy metal concentrations are spatially heterogeneous and strongly influenced by soil physicochemical properties. This study aimed to identify the key physicochemical properties affecting the performance of UAV-based hyperspectral models for estimating Zn, As, Cu, and Pb concentrations in soil. Soil pH, soil organic matter (SOM), and soil moisture content (SMC) were combined with spectral features to construct different input scenarios. The best-performing model for each metal was subsequently interpreted using Shapley additive explanations (SHAP). The predictive accuracy of the models for Zn, As, Cu, and Pb improved by 10
Agriculture 4.0 technologies such as automated machinery, precision farming, and data-driven decision-making can today help farmers manage the challenges of economic instability and climate adaptation. Yet uptake of on-farm technology is often variable, with reported barriers related to return on investment and technological integration. However, underlying these concerns can often be found a more human-centred issue. A lack of access to trusted information and subsequent subjective feelings of knowledgeability can hinder farmers from even beginning to explore pathways to adoption. This study used both quantitative and qualitative data to assess current levels of knowledge of automated technology amongst Australian grain growers and other rural stakeholders (N = 204). Findings revealed a picture in which self-reported knowledge was low and peers were cited as the most trusted sources of information. Quantitative analysis revealed links between knowledge and factors critical to adoption (such as intention), as well as factors impacting adoption-related decision-making (such as perceived usefulness). Indeed, in assessing which factors accounted for most variance in farmers’ intention engage with technology, it was a farmer’s sense of not being able to stay informed that had the most impact. These findings are important for understanding how to support the grains industry. Moreover, unlike some of the agricultural industry’s more intransigent challenges, building reliable and trustworthy knowledge networks represents an actionable and achievable industry goal.
Unmanned aerial vehicle (UAV)-based remote sensing can support non-destructive and spatial pasture monitoring, particularly in heterogeneous tropical systems. This study assessed the potential of UAV-derived vegetation indices (VIs) from the visible and near-infrared (NIR) spectrum to estimate canopy height and leaf area index (LAI) of ‘Tifton 85’ bermudagrass under conditions characterized with high spatial variability. Mean canopy height and LAI were measured at 55 georeferenced sampling points during five field campaigns between May 15 and June 22, 2023. Multispectral UAV imagery was used to derive 10 VIs. Relationships between VIs and observed variables were assessed using Spearman’s correlation coefficient (Rs). Simple nonlinear models were fitted and assessed using the coefficient of determination (R²) and the standard error of estimate (SEE), and tested with an independent dataset using R², root mean square error (RMSE), and the modified Willmott index (dmw). Near-infrared (NIR)-based indices showed the strongest relationships with pasture parameters. Similar to LAI, height was strongly correlated with some VIs (Rs > 0.80) and good fit to the nonlinear models (R² > 0.64). However, height model performance testing was poor (R² < 0.25, dmw < 0.49, and nRMSE > 40
Rice straw is abundant in yield, and utilizing straw as animal feed is a vital strategy for resource recovery. Among the factors affecting its feed value, moisture content is a particularly critical parameter. In this study, a rapid image-based method is proposed for detecting the moisture content of rice straw. Images of straw samples with moisture levels ranging from 0
Purpose. The study aims to substantiate the effectiveness of integrating Earth remote sensing (ERS) and geographic information systems (GIS) for implementing the principles of precision agriculture. The main objective is to analyze the potential of multispectral satellite data for assessing the spatial heterogeneity of agroecosystems, identifying vegetation stress conditions, and supporting decision-making in agricultural production management. Methods. The methodological basis of the study includes Earth remote sensing techniques, spectral analysis, and geoinformation modelling. Sentinel-2 multispectral imagery was used to calculate vegetation indices (NDVI, EVI, SAVI). Data processing involved atmospheric and radiometric correction, spatial classification of vegetation, integration of raster and vector data within a GIS environment, and spatial analysis methods (overlay analysis, interpolation, clustering) for delineating agro-technological zones. Results. The study demonstrated that the use of satellite data combined with GIS enables effective identification of intra-field variability in crop conditions and differentiation of zones with varying levels of bioproductivity. A stable correlation was found between vegetation indices and soil agrophysical properties. The implementation of a zone-based fertilization approach resulted in a reduction of mineral fertilizer use by 12–18 %, an increase in crop yield by 10–15 % in low-productivity areas, and a reduction in yield variability by 20–25 %. An environmental benefit was also observed through reduced risks of excessive nitrogen application. Practical significance. The results confirm the feasibility of implementing integrated GIS–RS approaches in precision farming systems. The proposed model enables optimization of resource use, improves the accuracy of agronomic monitoring, reduces the need for field surveys, and supports a transition toward spatially differentiated agroecosystem management. The obtained results can be applied in the development of decision support systems and digital agricultural platforms
Purpose. The study aims to substantiate and develop a methodology for operational monitoring of agricultural lands in the Rivne region using open data from Sentinel-1 and Sentinel-2 satellites. The main objective is to create an algorithm for assessing crop development dynamics and identifying soil degradation processes to support precise management decisions. Methods. The information base of the study consists of multispectral imagery from the Sentinel-2 satellite (Copernicus program), processed using the Google Earth Engine cloud platform and the Copernicus Browser service. Spectral analysis with NDVI (Normalized Difference Vegetation Index) calculation was applied to monitor winter crops during the 2024/2025 growing season on a field covering 56.8 hectares. The methodology involved a comparative analysis of Level-1C and Level-2A (Bottom-of-Atmosphere) processing levels to evaluate the impact of atmospheric distortions on the final indicators. Results. The constructed NDVI time series allowed for a clear identification of the key stages of winter wheat development: the active growth phase in April-May with values of 0.7–0.8, and a sharp decline to 0.2–0.3 during ripening and harvesting. It was established that Level-2A products provide higher accuracy for precision agriculture tasks due to atmospheric correction, although their absolute values are slightly lower compared to L1C. It was also revealed that the phenological stage of crop development significantly affects the accuracy of assessments, as changes in plant morphological features can distort index values. Practical significance. The proposed methodology enables continuous real-time monitoring of biomass status and plant health, minimizing the resources spent on visual field inspections. The results of the work can be integrated into decision support systems for agricultural enterprises for variable rate fertilization, irrigation optimization, and yield forecasting. A promising direction is identified in combining Sentinel- 2 optical data with Sentinel-1 radar data for operations under high cloud cover conditions typical of the Polissya region.
This paper presents an approach to crop condition assessment based on multi-index analysis using fuzzy logic. The aim of the study is to develop a methodology for integrated evaluation of agroecosystem conditions using spectral indices NDVI, NDRE, NDWI, and SAVI, which enables consideration of various physiological aspects of plant development. Proposed methodology is based on the processing of multispectral remote sensing data, calculation of vegetation indices, their normalization, and subsequent fuzzification using membership functions. The integration of indicators is performed through a fuzzy rule base followed by defuzzification to obtain a generalized assessment of field condition. An algorithm has been developed that includes the stages of index calculation, fuzzification, application of fuzzy rules, and defuzzification. The results demonstrate the capability of forming an integral assessment of vegetation condition considering biomass, chlorophyll activity, water balance, and soil background effects. The resulting generalized model allows for the identification of spatial variability within the field and the delineation of zones with different levels of vegetation development. Practical significance lies in the applicability of the proposed methodology in precision agriculture systems for decision support in crop management, optimization of input application, and improvement of agricultural production efficiency. The methodology is universal and can be implemented both in Python-based environments and in modern geographic information systems.
