Search Results (2,269)

Apple snail infestation poses a persistent threat to rice production in open-field environments, where long-term coexistence with this species is unavoidable. This study presents a drone-based precision control approach that integrates high-resolution micro-topographic mapping with site-specific pesticide application. A lightweight mapping unmanned aerial vehicle was deployed to produce centimeter-level microtopographic data across paddy fields, facilitating the identification of deep-water areas preferred by apple snails. From these elevation-derived water risk patterns, prescription maps were generated to guide downstream management decisions, and agricultural drones equipped for granular application subsequently performed targeted pesticide delivery only in these high-risk areas. Over 2 years of field experiments, the proposed method achieved rice yields comparable to those under conventional management while reducing pesticide use by 44.1–63.0%, with lower estimated crop damage in regions with high apple snail occurrence. Designed with robustness and scalability in mind, the system demonstrated considerable potential for practical implementation in general farming households and broader applications in precision pest management. Full article

Tropical estuarine hydrodynamic processes are governed by complex interactions between tides, monsoons, and fluvial runoff. To obtain long-term (≥30 years) hydrodynamic conditions of the Mekong River Estuary, this study established a Finite Volume Coastal Ocean Model (FVCOM) coupled with validated Weather Research and Forecast (WRF) wind forcing for a 32-year (1988–2019) high-resolution simulation. Validation against in situ observations confirms the model’s robustness. Temporal–spatial patterns of water level and current were analyzed, and extreme parameters for 1–100 year return periods were derived via the Pearson-III probability distribution. Results indicate the study area is a mesotidal environment (tidal range = 3.58 m) dominated by SSE-NNW reciprocating tidal currents. Relative to Vietnam’s national elevation datum, 100-year return period extreme high/low water levels are 2.15 m and −2.03 m, with a maximum storm surge setup of 2.09 m. The 100-year return period maximum current velocity reaches 4.58 m/s (A21 station), and Mekong River runoff exerts a negligible influence (<5% velocity change). This study provides high-precision baseline data for offshore wind farm engineering and disaster risk assessment, offering a methodological reference for tropical estuarine hydrodynamic simulations. Full article

Timely and accurate detection of sow oestrus is crucial for enhancing reproductive efficiency and reducing non-productive days (NPDs) in large-scale pig farms. However, traditional manual observation is labour-intensive and subjective, while cloud-based deep learning solutions face challenges such as high latency and privacy risks when applied in intensive housing environments. This study developed an edge-intelligent monitoring system that integrates deep temporal modelling with sound source localisation technology. A three-stage hierarchical screening strategy was utilised to select and deploy a lightweight Stacked-LSTM model on the resource-constrained ESP32-S3 hardware platform. This model was trained and calibrated using a high-quality acoustic dataset validated against serum reproductive hormones, specifically follicle-stimulating hormone (FSH), luteinising hormone (LH), and progesterone ($P_4$). Experimental results demonstrate that the optimised model achieved a classification accuracy of 96.17%, with an inference latency of only 41 ms, thereby fully satisfying the stringent real-time monitoring requirements while maintaining a minimal memory footprint. Furthermore, the system integrates a localisation algorithm based on Generalised Cross-Correlation with Phase Transform (GCC-PHAT). Through spatial geometric modelling, the system successfully implements the functional mapping of vocalisation events to individual gestation stalls (Stall IDs). Laboratory pressure tests validated the robustness and low-cost deployment advantages of the “edge recognition–cloud synchronization” architecture, providing a reliable technical framework for the precision management of smart livestock farming. Full article

The quality of wind power output data directly impacts the assessment of wind farm operational status and the accuracy of power forecasting models. However, due to factors such as sensor precision, communication interference, and the complex harbor environment, raw data collected from port-area wind turbines often contain noise, outliers, and missing values. Without effective cleaning, the resulting power curves can be distorted, reducing the generalization capability of predictive models. To overcome the limitations of traditional outlier detection methods in terms of adaptability and robustness, this study proposes a two-stage port-area wind power data cleaning approach based on dynamic interquartile range and an improved Sigmoid function fitting. In the first stage, an adaptive binning and density-weighting mechanism dynamically expands the interquartile range to identify and remove local outliers across different wind speed intervals. In the second stage, the cleaned wind speed–power data are subjected to secondary fitting and residual analysis using an improved Sigmoid model to detect hidden anomalies and boundary-type outliers. Using measured data from the #1 WT in the Chuanshan Port area as a case study, the experimental results demonstrate that the proposed method achieves high data retention while outperforming the conventional interquartile range, density-based spatial clustering of applications with noise and isolation forest algorithms in terms of the Pearson correlation coefficient (r = 0.93) and the coefficient of determination (R² = 0.89), with mean squared error and root mean squared error reduced to 446.39 kW and 545.58 kW, respectively. The findings verify the efficiency, stability, and practical feasibility of the method for port-area wind power data cleaning, providing a reliable data foundation for wind power forecasting and operational optimization in port environments. Full article

The advent of smart farms, enabled by information and communication technologies (ICT) and the Internet of Things (IoT), has improved productivity and sustainable agriculture. However, the large-scale implementation of smart farms is currently hampered by physical constraints. These constraints have led to the concept of open-field smart farming as a viable alternative. In this paradigm, data from unmanned aerial vehicles (UAVs) play a central role in effective and sustainable agricultural management. The quantitative analysis of such data requires highly reliable technological solutions. The objective of this study is to conduct a comparative analysis of image binarization algorithms for UAV-based soybean canopy extraction across growth stages and to contribute to the development of an image labeling methodology. UAVs were used to capture images of soybean fields at different growth stages, and a comparative analysis was performed using binarization image algorithms. The performance of each algorithm was evaluated using Normalized Cross Correlation (NCC) and Mean Absolute Error (MAE). The results indicate that the Excess Green (ExG) and Excess Green minus Excess Red (ExGR) vegetation indices provide accurate and stable soybean canopy extraction across growth stages when combined with Adaptive and Otsu binarization algorithms. These indices are particularly suitable for extracting soybean canopy from UAV-based data, thereby expanding the scope of precision analysis in the agricultural sector and providing data for advancing precision agriculture technology. This study contributes to the standardization and efficient use of UAV-based agricultural data processing. However, since manual weeding was performed prior to image acquisition to ensure that only soybean plants were present, reflecting standard agricultural practices in South Korea, additional validation would be required for application in fields where weeds are naturally present. Full article

In recirculating aquaculture systems, fish detection is an essential component for maintaining effective farming operations. The availability of high-quality fish datasets is limited because of the richness of fish species, and the annotation of large-scale data, which is used to train models, is often labor-intensive and time-consuming. The presence of different fish species across batches introduces further challenges for consistent detection performance. This work introduces a few-shot learning approach for fish detection, utilizing a customized dataset as novel classes and the Fish4Knowledge dataset for base classes, thereby establishing a framework that enhances adaptability in data-scarce scenarios. Within the model architecture, multi-scale feature extraction is enhanced through an attention mechanism, which is integrated as a dedicated module to strengthen representation learning, thus enhancing the model’s capability to differentiate visually similar fish species. Two distinct customized fish datasets are employed to evaluate the robustness of the proposed method. Experimental results show that the proposed model performs competitively against TFA, Meta-RCNN, and VFA. In the base-training phase, it achieves a mAP of 0.775, slightly surpassing VFA, while in the 1-shot, 5-shot, and 10-shot fine-tuning settings, it obtains mAP values of 0.152, 0.247, and 0.265, respectively. A similar trend is observed on a subset of black fish, with mAP scores of 0.169, 0.253, and 0.286 in the corresponding few-shot settings. These results indicate that the proposed approach can maintain relatively stable detection accuracy and adaptability across different fish batches, offering a practical solution for fish detection tasks in aquaculture when annotated data is scarce. To further demonstrate the efficacy and practical utility of the proposed methodology, a case study in fish farming confirms that the enhanced model achieves consistent and precise detection across diverse fish species, even when trained with limited annotated data. Full article

Vertical farming requires precise environmental control, yet long-term multivariable analyses linking environmental dynamics and crop growth remain limited. This study analyzes a two-year operational dataset from a commercial vertical farm in South Korea to evaluate the suitability of advanced artificial intelligence models for harvest yield prediction. Conventional machine learning models and recent deep learning architectures were systematically benchmarked under identical conditions. Among them, the patch-based Transformer model achieved the highest predictive accuracy (R² = 0.942; RMSE = 5.81 g per plant). The variable-importance analysis revealed that daily light integral and CO₂ concentration were the dominant drivers of harvest yield variability, jointly accounting for more than 76% of total contribution. These findings demonstrate the effectiveness of Transformer-based architectures for long-term multivariate time series modeling and provide actionable insights for the data-driven optimization of vertical farming systems. Full article

Precision agriculture employs state-of-the-art technologies to improve the economic viability, sustainability, and efficiency of agricultural practices. This paper offers a thorough review of precision agriculture, with an emphasis on real-time systems as a foundation for understanding the integration and impact of major technologies. We examine technologies such as digital twins, mobile applications, autonomous systems, location-aware technologies, edge computing, and Wireless Sensor Networks (WSN) that are revolutionizing agricultural processes. We also discuss the potential of other sensing techniques to enhance precision farming, including image analysis, sensory and chemical analysis, and physical state detection. Additionally, the roles that data transmission protocols, artificial intelligence (AI), and machine learning play in maximizing real-time data processing and decision-making are examined. We emphasize the main challenges and limitations in precision agriculture, such as data interoperability, scalability, and system integration. With a focus on market trends and local issues, we examine how AI, real-time systems, sensor technologies, and financial constraints impact the growth of precision agriculture. These advancements have an impact on precise monitoring, post-harvest management, and human health. Lastly, we provide suggestions for successful integration and future developments in precision agriculture, emphasizing design, engineering, and creative approaches to assist the field’s ongoing development. Full article

The management of specialty coffee production represents a complex dynamical process characterized by highly nonlinear interconnections between environmental variables, agronomic practices, and chemical compositions. Traditionally, the classification of specialty coffee relies on sensory evaluations conducted by highly certified coffee experts named Q-Graders, using a strict, standardized Specialty Coffee Association (SCA) protocol. However, scientific methods that generate spectral fingerprints provide a more reliable guarantee of quality while also ensuring traceability to the farm of origin. Panamanian Geisha coffee is one of the world’s most expensive award-winning microlots, frequently exceeding 1000 American dollars per pound, with a record-breaking price of over 30,000 American dollars per kilogram in 2025. This research presents an integrated framework that combines Precision Agriculture Management Systems (PAMSs) and a traceability architecture that facilitates the collection of georeferenced coffee bean samples using a mobile application (apps), while preserving the coffee varieties and geographical origin necessary for the subsequent identification of the spectral fingerprint by chemical specialists in their laboratory. A mathematical model is introduced to formally characterize the mobile application’s behavior, distributed structure, and inherent constraints. Serving as a mathematical blueprint, this model identifies critical influencing factors and establishes strategic assumptions to distill complex real-world variables into a rigorous, manageable framework. Large-scale experiments conducted across more than 820 coffee farms in Chiriquí, Panama, demonstrate that the proposed decentralized architecture effectively coordinates the acquisition and synchronization of georeferenced chemical data. The decentralized architecture of the mobile application utilizes private blockchain technology to facilitate autonomous operations, effectively decoupling the system from central authorities to ensure functional continuity in environments characterized by intermittent connectivity. Full article

The adoption of Internet of Things (IoT) technologies has significantly enhanced precision farming by enabling continuous environmental monitoring and data-driven agricultural management. However, the increasing reliance on distributed sensor networks introduces critical challenges, including sensor faults, data anomalies, and cyber-physical security threats, which can undermine system reliability and decision accuracy. This study proposes an IoT-driven anomaly detection framework for smart agriculture that integrates a Convolutional Neural Network (CNN) optimized using a hybrid Particle Swarm Optimization and Genetic Algorithm (PSO–GA). The CNN learns complex spatio-temporal patterns from multivariate sensor data, while the PSO–GA strategy automatically tunes CNN hyperparameters to improve detection accuracy and model stability. To enhance adaptability under dynamic agricultural conditions, the proposed framework incorporates an online learning mechanism that incrementally updates the CNN model using newly arriving sensor data, enabling continuous adaptation to environmental changes and concept drift without full model retraining. Experiments conducted on a publicly available smart agriculture dataset demonstrate that the proposed CNN–PSO–GA framework achieves an accuracy of 74%, precision of 74%, recall of 100%, and an F1-score of 85%, outperforming baseline methods such as One-Class Support Vector Machine and Isolation Forest, particularly in reducing missed anomaly events. The results confirm the robustness, adaptability, and reliability of the proposed approach. Overall, the framework provides a practical and scalable solution for enhancing security, resilience, and operational effectiveness in precision farming systems. Full article

The integrated crop–livestock farming model not only enhances resource utilization efficiency and reduces environmental pollution but also serves as a vital pathway for promoting sustainable agricultural development. System dynamics models enable in-depth analysis of dynamic system changes. Therefore, in this study, we constructed a comprehensive system dynamics model for the combination of farming and raising in Jilin Province, and used Vensim-PLE software(Version 6.3) to simulate and predict the dynamic changes of the agricultural structure and the future development trend based on the data of the planting and raising industry from 2006 to 2021. The results of the study show that: (1) It is expected that by 2035, the grain crop production, cash crop production, hog and beef cattle output, as well as the utilization of straw and livestock and poultry manure in Jilin Province will increase significantly. And the livestock and poultry manure load warning value reaches 0.35, which does not pose a threat to the environment. (2) By setting the simulation structure optimization analysis, it is found that the integrated crop-livestock systems performs the best in terms of economic, social, ecological and environmental benefits. (3) In 2035, Jilin Province should control the total crop planting area within 7250 thousand hectares, the number of hog output should not exceed 25 million, and the number of beef cattle slaughtered should not exceed 4.25 million, so as to ensure that the early warning value of livestock and poultry fecal matter loading is lower than 0.4, and to achieve the balance of farming and maximize economic benefits. Finally, this paper proposes policy recommendations to optimize crop structures, develop precision farming and aquaculture technologies, establish resource recycling projects, and strengthen policy support and technology promotion. Full article

Accurate plant classification based on deep learning is important for precision agriculture, such as weed control, crop monitoring, and smart farming systems. The accuracies of deep learning models rely on datasets. Although many datasets have been proposed in recent decades, they have the common limitations in terms of scale, less environmental diversity, and challenges of data integration. To solve these problems, in this paper, we introduce a new dataset named OpenPlant, which is a large-scale and open dataset containing 635,176 RGB images across 1167 plant species. OpenPlant includes diverse growth stages of plants, plant structures, and environmental conditions, and its annotations were carefully verified to ensure quality. The proposed OpenPlant can be a benchmark for agricultural plant classification. In this paper, we benchmarked 10 widely used convolutional neural networks (CNNs), 6 vision transformers (ViTs), and 12 vision–language models (VLMs) to provide a comprehensive evaluation. The OpenPlant dataset offers a comprehensive benchmark for agricultural research using deep learning and the results provide insights into future directions. Full article

The CSN2 gene encoding β-casein has gained increasing attention in dairy cattle breeding due to the global adoption of A2-oriented selection strategies. However, robust large-scale evidence assessing potential unintended effects on functional traits, particularly fertility, under intensive commercial conditions remains limited. This study evaluated whether selection for the CSN2 A2 β-casein variant is associated with biologically relevant differences in fertility traits in Holstein cows. Reproductive and genomic data from 7826 lactation records of 2773 Holstein cows collected between 2022 and 2025 in a large commercial dairy herd were analyzed. Fertility indicators included days open, number of services per conception, calving interval, first-service conception rate, and pregnancy by 100 days in milk. Mixed-effects models accounting for repeated lactations and cow- and sire-level clustering were applied, and predefined equivalence margins were used to distinguish statistical non-significance from biological irrelevance. Across all evaluated fertility traits, differences among CSN2 genotypes (A1A1, A1A2, and A2A2) were consistently small, biologically negligible, and well within predefined equivalence margins. Differences in days open were within ±2 days, and effect sizes for count and binary traits were close to unity. Parity and calving year significantly influenced reproductive performance, whereas no CSN2 genotype × parity interactions were detected. These findings indicate that selection for the CSN2 A2 β-casein variant does not compromise reproductive performance under intensive commercial management conditions. From a breeding and industry perspective, the results support the implementation of A2-oriented selection strategies without biologically meaningful adverse effects on fertility. Full article

Precise monitoring of sow behaviour is essential for enhancing animal welfare and production efficiency in precision husbandry. This study proposes an improved RT-DETR model to address real-time detection challenges in complex farming environments. By integrating innovative multi-scale feature fusion and lightweight attention mechanisms, the model achieves high-precision detection of four key postures (standing, sitting, sternal recumbency, and lateral recumbency). Experimental results show that the model attains an mAP@0.5 of 96.6% and a processing speed of 56 FPS, significantly outperforming existing methods. Furthermore, a Unity3D-based digital twin system was constructed to enable real-time bidirectional mapping, achieving a low latency of 320 ms. This system proposes a potential technical framework for intelligent pig farm management, providing a reliable tool for automated welfare assessment and operational decision support. Full article

Accurate feeding-state monitoring is essential for improving feeding management, reducing feed waste, and supporting water quality and fish welfare in aquaculture. However, existing vision-based methods often rely on subjective labels or computationally expensive temporal models, which limits practical on-farm deployment. Here, we propose an objective, edge-deployable framework for motion-driven feeding-state quantification and binary feeding/non-feeding recognition from top-view videos. The framework integrates frame-pair dense optical-flow encoding with a lightweight network (EfficientFeedingNet) to enable real-time deployment. Using an optical-flow-derived motion-intensity signal (V-Value), we automatically delineate feeding-response intervals and construct a perception-based dataset (Perceptual Dataset) with reproducible binary labels, alongside an observer-labeled Intuitive Dataset. Across representative backbones, models trained on the Perceptual Dataset achieve >90% test accuracy and improve over the Intuitive Dataset by 13.13–18.46 percentage points. The proposed EfficientFeedingNet attains 96.53% test accuracy while remaining lightweight for edge deployment; on a Jetson Orin NX, it runs at 7.0 ms per image (143.24 fps). Overall, the proposed framework provides a practical basis for timely, data-driven feeding decisions in precision aquaculture. Full article