Search Results (35,282)

Thermal infrared (TIR) whole-sky imaging enables continuous cloud observations during both daytime and nighttime, making it an important tool for ground-based cloud monitoring. However, existing infrared cloud detection methods mainly rely on manually selected thresholds or handcrafted image features, which typically require instrument-specific calibration and exhibit limited generalization capability. Moreover, wide-field imaging covers not only the zenith region but also the peripheral areas at large zenith angles, which are susceptible to the combined effects of strong atmospheric background emission, thermal radiation contamination from the surface and surrounding environment, and imaging geometric distortion, resulting in reduced cloud–background contrast and making cloud segmentation considerably more challenging than in conventional infrared sky imagery. To address these limitations, this study proposes a deep-learning framework for automatic cloud detection and cloud cover retrieval from wide-field TIR whole-sky images. The framework adopts a two-stage strategy in which infrared images are first classified as clear-sky or cloudy, followed by semantic segmentation of cloudy images for cloud cover estimation. To support model training and evaluation, a dataset containing 3000 wide-field TIR whole-sky images was established using a semi-automatic labeling procedure. Experimental results demonstrate that the proposed method effectively identifies cloud structures, including those located near the edge of wide-field whole-sky images. The classifier achieved an overall accuracy of 98.44%, the segmentation network attained a mean pixel accuracy of 90.44%, and the derived cloud cover showed excellent agreement with the manually annotated reference masks with a correlation coefficient of 0.92. Ablation studies further validate the effectiveness of the proposed deep-learning framework in improving the segmentation of complex cloud structures. The truncated EfficientNet-B0 encoder yields the most pronounced contribution, while the ASPP module improves the delineation of complex sky and cloud structures. When applied to the observations from the ground-based TIR all-sky camera, the retrieved cloud cover agreed well with the FY4B/AGRI cloud mask product, with a correlation coefficient ranging from 0.81 to 0.89 across the four seasonal months in 2023. These results prove that the proposed framework provides an effective and robust solution for long-term ground-based cloud monitoring using wide-field TIR whole-sky imagery. It also offers a practical foundation for applications such as cloud climatology, solar radiation assessment, and atmospheric observation. Full article

Low Earth orbit (LEO) satellite networks exhibit highly dynamic topologies, making it challenging to accurately identify critical inter-satellite links using conventional static graph methods. Existing approaches predominantly rely on delay or betweenness centrality as evaluation criteria and fail to account for temporal connectivity as a fundamental prerequisite for communication. This paper proposes the Evaluation of Temporal-Aware Path Connectivity (eTAPC), a method based on temporal connectivity for identifying critical links in LEO satellite networks. An event-driven slot partitioning strategy is adopted to construct a time-extended graph. Three complementary metrics are proposed, activity frequency, burstiness coefficient, and von Neumann entropy change, to characterize link importance from temporal coverage, fluctuation patterns, and topological bridging. These metrics are fused via the TOPSIS method with a predefined weight vector to produce a comprehensive ranking of link criticality. Experiments are conducted on real TLE data from both dense and sparse LEO constellations to validate the effectiveness of eTAPC. The results show that eTAPC outperforms existing methods on the dense Starlink constellation and achieves competitive performance on the sparse Iridium constellation. Overall, eTAPC provides a lightweight and effective solution for critical link identification in large LEO constellations, with low computational complexity. Full article

Smart agriculture and smart breeding are evolving from isolated sensing toward geographically distributed, long-term, data-driven closed-loop management. Dispersed farms and breeding sites, UAV and remote-sensing phenotyping, real-time field control, edge inference, and digital-twin synchronization impose heterogeneous demands on coverage, latency, bandwidth, reliability, computing, and energy. Space–air–ground integrated networks (SAGINs) combine satellites, unmanned aerial vehicles/high-altitude platforms, and terrestrial networks for wide-area coverage and elastic access, but introduce dynamic topologies, heterogeneous multi-domain resources, and complex cross-layer orchestration. Focusing on intelligent SAGIN resource management enabled by software-defined networking (SDN) and network function virtualization (NFV), this review examines controller placement, NFV/service function chain orchestration, SDN/NFV cooperation, learning-driven optimization, and security mechanisms. Representative studies are compared by objectives, decision variables, mechanisms, applicability boundaries, and engineering costs. These mechanisms are connected to smart agriculture and smart breeding through ubiquitous connectivity, edge computing, task offloading, phenotyping, and digital-twin closed loops, while distinguishing experimentally supported agricultural evidence from SAGIN-oriented architectural inference. Future research should move beyond single-metric optimization toward joint evaluation of state freshness, decision latency, reconfiguration cost, learning and security overhead, and agronomic outcomes to improve deployability, robustness, and verifiability of cross-domain resource management under real deployment conditions across heterogeneous agricultural environments. Full article

In this study, a machine learning model for the anisotropic mechanical behavior of AA7075-T6 was successfully developed by combining classical plasticity theory with neural network algorithms. First, a yield criterion incorporating strength asymmetry and orientation dependence was constructed based on a fourth-order linear transformation of invariants, and a stress update algorithm was implemented using the return mapping and implicit integration schemes. Subsequently, a dataset comprising experimental results from tension, compression, and shear tests at multiple orientations, as well as theoretically generated data from additional strain paths, was established to train a genetic algorithm-optimized two-hidden-layer neural network. Plastic-stage assessments indicated that, for the equal-biaxial path, the RMSE values of the stress–strain curves predicted by the machine learning model relative to the constitutive implementation were 31.87 and 39.24 MPa. It should be noted that the stress–strain response under equal-biaxial loading represents an additional prediction case. The trained model exhibited satisfactory predictive performance when compared against experimental data for tension, compression, and shear and was capable of reproducing the theoretical stress paths derived from the classical constitutive model. This approach leverages the physical interpretability of conventional constitutive modeling and the high efficiency of data-driven methods, providing a viable solution for efficiently predicting the anisotropic mechanical response of metallic materials under complex stress states. Full article

Livestock farms are a distinctive type of agricultural ecosystem wherein farmed animals interact continuously with other living organisms and with the physical environment of the farm under the influence of human management. The objectives of this paper are the proposal of a framework in which small ruminant farms are considered as managed ecosystems, the identification, definition and characterization of the principal biotic components and abiotic environmental compartments, the examination of interactions occurring within and between these components and the description of the role of health management in modifying interactions within these systems. Within the proposed approach, the ecosystem of small ruminant farms can be considered to be characterized by five principal elements: components and compartments, interactions and flows and their directionality, inputs and outputs crossing the physical or functional boundaries of the system, and human management and feedback. Small ruminant farms can function as networks of ecosystem interactions, in which complex interactions occur among and between their biotic components and abiotic environmental compartments. In brief, the biotic components of the farm ecosystem include the livestock (sheep, goats), other domestic or synanthropic animals on farms, wildlife near farms, people working on farms or visiting, vegetation and other primary producers, pathogens, arthropods and other biotic communities. The abiotic environmental compartments include animal housing and associated areas and materials, farm systems, machinery and light equipment, feed and water, climate variables and farm waste. Various interactions connect those components. Health management acts upon ecosystem components and their interactions and is a principal means through which humans deliberately modify the farm ecosystem; it is based on human decisions, which influence multiple components of the ecosystem and direct the interactions within the system. Health management can modify interactions throughout the farm ecosystem, with effects that may extend beyond their immediately intended targets. The consideration of these wider relationships can support better-informed and integrated health management of sheep and goat farms and their animal populations. Full article

Underwater sensor networks (USNs) are essential for marine exploration and monitoring, yet their performance is limited by transmission reliability in complex underwater environments. Underwater acoustic (UWA) transmission offers a practical solution, with orthogonal frequency division multiplexing (OFDM) extensively adopted for its high data rate and multiple access capability. However, OFDM is extremely vulnerable to Doppler-induced distortions, which degrade demodulation performance. Conventional cross-ambiguity function (CAF) methods estimate Doppler through signal-level correlation, but achieving high accuracy requires long training sequences, thus incurring frame overhead and reducing effective data rates. To address this issue, a novel bit-level correlation (BLC) Doppler estimation algorithm enables accurate estimation with diminished training overhead. A tailored OFDM frame uses the first two OFDM symbols, modulated with M-sequences, as training sequences. Demodulated bits are correlated with a local M-sequence via vector inner-product computation, converting correlation from the signal level to the bit level. Numerical simulations under three Doppler scales demonstrate the effectiveness of the BLC algorithm. By correlating the demodulated training bits with a local M-sequence reference, the proposed algorithm provides a favorable trade-off between training overhead and estimation accuracy under the tested simulation conditions. Full article

To clarify the controlling effects of the well deviation angle and azimuth angle on the breakdown pressure, propagation morphology, and fracture-network complexity of hydraulic fractures in tight sandstone, true triaxial hydraulic fracturing physical simulations were systematically conducted on Chang 7 Member sandstone from Yanchuan County, Ordos Basin, under different well deviation and azimuth angles. By combining injection-pressure monitoring, surface fracture-morphology observation, and three-dimensional laser scanning, the breakdown pressure, propagation path, surface roughness, fractal dimension, and overall complexity of the fractures were quantitatively analyzed. The results show that, at an azimuth angle of 90°, the breakdown pressure of the sandstone generally decreases as the well deviation angle increases from 0° to 90°, dropping from 19.125 MPa to 13.569 MPa, indicating that horizontal wells are easier to fracture. At a well deviation of 60°, the fracture is more prone to deflect and communicate with natural weak planes, yielding the highest overall complexity (f = 1.629). For horizontal wells under normal-faulting stress, the breakdown pressure decreases as the azimuth angle increases; the lowest breakdown pressure (12.933 MPa) is obtained when the wellbore is drilled along the maximum horizontal principal stress (σ_H), and the highest (18.310 MPa) when parallel to the minimum horizontal principal stress (σ_h). When the azimuth angle is 30°, both the fracture-surface roughness (S_a = 2.037 mm, S_q = 2.691 mm) and the overall complexity (f = 1.679) reach their maxima, which is most favorable for forming tortuous, rough, and complex fracture networks. The fractal dimension of the fractures varies little across the tested conditions (2.0339–2.1379), indicating that it is mainly controlled by the intrinsic heterogeneity of the rock. The research results can provide an experimental basis for the optimization of horizontal-well trajectories and fracturing-parameter design in tight sandstone reservoirs. Full article

Drug-induced pseudoallergic reactions are rapid in onset and complex in mechanism, with current clinical safety assessments lacking precise molecular underpinnings. Cells were treated with pseudoallergen-positive compounds. Cell degranulation was verified by neutral red staining and ELISA, followed by high-throughput RNA sequencing. Key signaling pathways were elucidated through Gene Ontology and Kyoto Encyclopedia of Genes and Genomes enrichment analyses. A pseudoallergic characteristic gene network was identified and constructed using WGCNA. Finally, two-sample MR analysis was performed utilizing GWAS data to evaluate the causal relationship between candidate genes and drug pseudoallergy, thereby identifying key genetic risk markers. The pseudoallergen-positive compounds significantly induced cell degranulation and extensive transcriptional regulation. Enrichment analyses indicated that MAPK, JAK-STAT, TNF, and PI3K-Akt signaling pathways play central roles in pseudoallergic reactions. WGCNA identified a pseudoallergic characteristic gene signature. Subsequent MR analysis identified EMP1 and KIF26A might be the key genetic risk markers associated with drug pseudoallergy. These two genes potentially increase the causal risk of drug-induced pseudoallergy by modulating membrane coating, vesicle formation, and related cellular transport processes. This study successfully established a mast cell characteristic gene signature for pseudoallergic reactions, elucidated the core transcriptional network, and identified key risk markers that may have a causal association with drug pseudoallergy. Full article

This paper investigates two integrable $(2+1)$-dimensional nonlinear evolution equations, namely the Kairat–II–X-extended (K–II–X-E) model and the Kairat–II-X (K–II–X)-type model, with the objective of deriving exact analytical soliton solutions. These models serve as fundamental prototypes for nonlinear wave dynamics and arise in a broad range of physical contexts, including fluid flows, optical signal propagation, plasma media, and selected biological systems. A $(G^{\prime }/G)$-expansion neural networks (ENNs) framework is developed by rigorously coupling the algebraic mechanism of the classical method with the symbolic neural networks. This unified strategy enables the systematic construction of exact analytical solutions directly from the governing equations. Using this framework, exact one- and two-soliton solutions are constructed, along with a broader class of nonlinear wave structures, demonstrating the capability of the method to capture diverse solution dynamics. The resulting solutions are expressed explicitly in trigonometric, hyperbolic, and rational forms. The dynamical features of the obtained solutions are further illustrated through 2D and 3D surface and polar plots, which reveal their localization properties, propagation characteristics, and interaction patterns in higher-dimensional settings. The results show that the $(G^{\prime }/G)$-ENNs method provides an efficient and systematic framework for constructing exact solutions of nonlinear evolution equations. This approach offers a promising analytical tool for investigating complex wave phenomena and underscores the potential of neural-network-assisted symbolic techniques in advancing research in nonlinear mathematical physics. Full article

Background: Alzheimer’s disease (AD) is a complex neurodegenerative disorder with a global impact, highlighting an urgent need for effective dietary therapeutic agents for the development of brain-healthy functional foods and medicines. This study focuses on the edible red alga Pterocladiella capillacea from Zhanjiang, China, investigating its anti-neuronal injury activity, active principles, and underlying mechanisms. Methods: Based on antioxidant and neuroprotection evaluations and column fractioning, bioactivity-coupled liquid chromatography-tandem mass spectrometry (Bio-LC-MS/MS) was employed to recognize the neuroprotective compounds, which were further annotated by comprehensive metabolomics analysis. The annotated compounds were validated by LC-MS/MS and bioassay. Computational prediction techniques were further used to reveal their mechanism of action. Results: The study demonstrated that P. capillacea extract and fractions exhibited significant antioxidant and anti-neuronal injury activities. Among the annotated and MS-validated potential compounds, erucamide standard exhibited the strongest neuroprotective effect, maintaining cell viability above 85% at low concentrations. Network pharmacology and molecular docking revealed that erucamide may exert neuroprotective effects by modulating the AKT1 and BCL2 targets. ADMET prediction suggested a low risk of central side effects. Conclusions: The neuroprotective potential of P. capillacea provides a primary foundation for new brain-healthy functional foods and drug development in future. Full article

Automatic row guidance is important for efficient and low-loss corn harvesting. However, complex field conditions challenge reliable inter-row perception, while many deep learning models remain computationally demanding. To address this, a lightweight semantic segmentation network is proposed. A dataset covering challenging field conditions was constructed. Based on DeepLabV3+, GhostNetV2 was adopted to reduce computational cost, an SP-ASPP was designed to enhance the representation of elongated inter-row structures, and BiFormer was introduced to strengthen long-range contextual modeling. We jointly exploited directional multi-scale context and sparse long-range interactions to preserve continuous row-space structures under occlusion and background interference while maintaining a lightweight architecture. A composite loss combining Focal, Dice, and Boundary losses was further employed to improve region completeness and boundary localization. A navigation-line extraction algorithm was then developed to generate stable guidance paths. After structured pruning and TensorRT FP16 optimization, the model achieved an mIoU of 82.28% at 53.1 FPS on the edge platform. The extracted navigation line yielded a mean absolute lateral error of 4.2 cm and a mean absolute heading error of 2.17°. These results demonstrate that the proposed method provides accurate real-time navigation perception on resource-constrained hardware, supporting low-cost vision-based corn harvester guidance. Full article

As a disruptive technology to overcome the bottlenecks of conventional mechanical excavation and blasting, microwave-assisted rock fragmentation demonstrates tremendous potential in deep geo-resource development and extreme-environment engineering. This paper presents a comprehensive review of the latest advances in this field, thoroughly dissecting the electromagnetic–thermal–mechanical–chemical (THMC) multi-field coupling mechanisms governing microwave–rock interactions. Particular emphasis is placed on elucidating the mechanisms of thermal stress fracturing, phase-transition expansion, and thermo-chemical coupling damage triggered by mineral dielectric heterogeneity. The nonlinear effects of microwave radiation parameters, intrinsic rock properties, and complex in-situ stress environments on fracturing efficiency are rigorously analyzed, and precision multi-scale characterization methodologies—from macroscopic mechanical degradation to microscopic fracture networks—are comprehensively summarized. Building upon this foundation, the paper critically evaluates the current engineering application status and electromagnetic safety shielding challenges of this technology in frontier scenarios, including microwave-assisted tunnel boring machines (TBM), intelligent mineral sorting, deep unconventional oil and gas fracturing, and space in-situ resource utilization (ISRU). Finally, the limitations of existing research regarding cross-scale effects, high-temperature and high-pressure in-situ testing, and dynamic intelligent control are identified, and future development pathways oriented toward multi-source data fusion and adaptive power modulation are prospected, aiming to provide robust theoretical support and forward-looking guidance for the cross-scale engineering transformation of microwave rock-breaking technology. Full article

This study explores the application of complex network theory to examine the temporal dynamics of groundwater levels in a region. It presents the first ever application of a coupled chaos theory–complex network framework to investigate the temporal dynamics and connectivity in groundwater levels in a region. For implementation, daily groundwater-level data observed over the period 2010–2014 from each of 71 wells across the United States are considered. Each well is considered as a network. In order to construct the temporal network from groundwater time series, basic concepts of chaos theory are coupled with those of complex network theory. First, the phase-space reconstruction technique is used to reconstruct the single-variable groundwater time series in a multi-dimensional phase space. Next, the optimum dimension required for the phase-space reconstruction is determined using the false nearest neighbour (FNN) algorithm. Each reconstructed vector in the phase space is considered as a node in the network, and the connections between the nodes (i.e., links) are identified based on the distance threshold between the reconstructed vectors. Four complex networks-based measures—degree centrality, betweenness centrality, closeness centrality, and clustering coefficient—are used to examine the properties of the groundwater-level network. The optimal embedding dimensions from the FNN method for the 71 groundwater level time series are found to range from 4 to 18, suggesting a wide range of complexity in the 71 groundwater level time series. However, a large majority (i.e., 62) of the time series have dimensions in the range of 4–10, suggesting low-to-medium-level complexity of the groundwater level dynamics. The results for the four network measures are degree centrality in the range 0.12–0.55, betweenness centrality in the range 604–4750, closeness centrality in the range 0.00009–0.00034, and clustering coefficient in the range 0.709–0.863. No region-specific patterns are observed in the four network measures, although, depending upon the measure, networks of wells in some states have higher/lower values and networks in some other states have a good mix of high and low values. Further, Spearman rank correlation analysis shows that the four network measures do not have significant relationships with key statistical characteristics (mean, standard deviation, and coefficient of variation) of the groundwater levels. This indicates that the differences in the network measures cannot be explained only by the basic statistical properties of the groundwater level records. Full article

Internet of Things devices introduce significant security challenges caused by their heterogeneous and resource-constrained nature in many sectors, such as healthcare, industry, education, and agriculture. Intrusion detection systems (IDSs) serve a key role in identifying malicious activities in such environments; traditional approaches cannot often capture dynamic interactions and temporal correlations in network traffic. In this research, we propose a novel hybrid IDS technique utilizing Long Short-Term Memory (LSTM) networks in conjunction with Lotka–Volterra (LV) dynamic modeling. The LSTM component is employed to learn temporal patterns and estimate system states from IoT traffic, while the LV model captures the dynamic interaction between normal and malicious behavior. We introduce a mathematically based decision mechanism on an anomaly score for effective classification. The model’s results on DataSense: CIC IIoT dataset 2025 achieve an accuracy of 99.85%, outperforming other models, and have a detection time of 47 ms and a reduced-complexity algorithm. These results highlight the effectiveness of combining artificial intelligence with dynamic system modeling for intrusion detection in IoT environments, providing a promising direction for future research in intelligent cybersecurity systems. Full article

Precise indoor thermal environment forecasting is critical for the energy-efficient operation and model predictive control of building heating, ventilation, and air-conditioning (HVAC) systems. However, conventional predictive models often struggle to untangle the complex spatiotemporal dynamics captured by sparse sensing networks. This study proposes a Spatio-Temporal Joint-Embedding Predictive Architecture (ST-JEPA) for indoor multi-node temperature and humidity prediction. By reformulating predictive representation learning from visual domains to structured sensor-node sequences, the proposed model explicitly integrates historical multi-node observations, spatial positional encoding, and concurrent HVAC operating states. The model was evaluated at prediction horizons from 30 s to 5 min using 3568 frames recorded over approximately 29.8 h by a nine-node sensor array in a single laboratory. Experimental results indicate that ST-JEPA achieves high short-term forecasting accuracy, with temperature root mean square errors (RMSE) of 0.0490 °C at 30 s and 0.0647 °C at 1 min. In the original ablation experiment, removing HVAC operating-state inputs increased temperature RMSE by up to 68.3%. Aggregate errors remained relatively stable under the tested 30–50% node-masking conditions, although additional single-node tests revealed location-dependent sensitivity. Comparisons with simpler baselines showed no consistent superiority across targets and horizons. ST-JEPA characterizes the short-term evolution of indoor thermal fields, providing a potential forecasting basis for HVAC feedforward regulation; its effects on energy consumption and thermal comfort remain to be evaluated. Full article