Search Results (20,568)

Early identification of thermal and electrical anomalies in grid-connected photovoltaic (PV) systems is becoming increasingly important to reduce energy losses, limit power quality (PQ) degradation, and avoid excessive operating stress on power electronic converters. Conventional electrical monitoring methods can provide overall performance information, but they are generally unable to detect and localize early-stage defects occurring at module or cell level. In this context, the present study proposes an integrated diagnostic framework that combines non-destructive infrared thermography (IRT) with advanced electrical signal processing techniques for PV condition monitoring. The proposed approach correlates thermographic information, capable of revealing defects such as hotspots, cell cracks, and bypass diode failures, with high-frequency electrical signal analysis based on frequency-domain and time–frequency methods, together with deep learning-driven thermographic segmentation. By associating thermal acquisitions with electrical PQ indicators, the framework enables the early detection of physical defects linked to inefficient Maximum Power Point Tracking (MPPT) operation and progressive degradation of PV system performance. The methodology was experimentally validated on a grid-connected photovoltaic installation under different fault conditions, including hotspots, bypass diode anomalies, and localized overheating effects, demonstrating the potential of the proposed approach for predictive maintenance and intelligent PV monitoring applications. The obtained results indicate that the proposed framework improves the reliability of photovoltaic fault detection by combining thermographic inspection with advanced electrical signal analysis and AI-based defect interpretation, thus supporting predictive maintenance strategies in smart PV infrastructures. The proposed approach demonstrates image segmentation capabilities, as evidenced by a precision (PA) of 96.88%, a mean IoU (mIoU) of 77.83% and a macro F1-score of 87.47%. The proposed framework maintained reduced computational requirements compatible with real-time monitoring applications. Full article

Background: Near-infrared spectroscopy (NIRS)-based wearable devices offer non-invasive, continuous monitoring of muscle oxygenation, providing direct microvascular and metabolic information that complements indirect indices of intensity such as heart rate and accelerometry. Their clinical applicability in chronic non-communicable diseases (NCDs) remains under active development. Methods: A structured narrative review was conducted in PubMed, Scopus, Web of Science, and IEEE Xplore (January 2010–January 2026) using pre-specified search strings combining NIRS, muscle oxygenation, SmO₂, StO₂, wearable, exercise intensity, ventilatory/lactate threshold, and individual chronic disease terms. Eligible studies addressed technical validation of wearable NIRS, NIRS-derived exercise intensity estimation, clinical applications in NCDs, or rehabilitation implementation. Evidence was synthesized thematically; quality of validation studies was appraised against AMSTAR-2-informed, COSMIN-informed, or Cochrane RoB-2 criteria. Results: Wearable continuous-wave NIRS shows acceptable concurrent validity with frequency-domain laboratory systems (r = 0.79; range 0.69–0.88; ±8% SmO₂ agreement in 95% of measurements) and good test–retest reliability for moderate-to-severe domains (ICC 0.72–0.91). NIRS-derived breakpoints align more reliably with the second ventilatory/lactate threshold (ICC = 0.80) than with the first (ICC = 0.53), constraining its use for prescribing lower-intensity domains. In chronic obstructive pulmonary disease, peripheral arterial disease, chronic respiratory failure and selected cardiovascular conditions, wearable NIRS detects disease-specific patterns of muscle deoxygenation and post-exercise reoxygenation that track responses to rehabilitation. Conclusions: Current evidence supports wearable NIRS as a complementary, intensity-aware monitoring tool—particularly for delineating the heavy/severe-intensity boundary and detecting peripheral metabolic limitations—rather than as a stand-alone replacement for ventilatory or lactate thresholds. Because much of the evidence derives from small, single-sex or athlete-only cohorts, these findings should be regarded as a promising basis requiring further validation in broader NCD populations. Implementation in NCDs requires standardized placement and calibration protocols, sex- and body composition-stratified reference values, motion-artifact mitigation, and adequately powered longitudinal trials in clinical populations. Full article

Sovereign credit ratings are central to a country’s access to international capital markets, yet the relative informational content of debt service obligations versus aggregate debt stock for rating outcomes remains empirically unsettled. This systematic review synthesises econometric evidence on both measures across 23 primary studies published between 1996 and 2024. The central message of this paper is that debt service indicators—capturing near-term liquidity and refinancing pressure—are at least as informative as, and on average more informative than, debt stock ratios for sovereign credit assessments, particularly in emerging-market contexts and ordered-response specifications. This finding holds across heterogeneous study designs and is confirmed by meta-regression analysis, which shows that debt service effects are significantly more negative than debt stock effects (β = −0.09, p = 0.004) after controlling for sample composition, model family, and rating agency. Emerging-market samples and ordered-response estimators yield stronger associations than advanced-economy samples and linear (OLS) specifications. No consistent differences across the major rating agencies are found once study-design moderators are included. Because primary studies differ in model families, samples, and variable construction, we emphasise transparent reporting, avoid over-interpreting pooled magnitudes, and focus on robust qualitative patterns and moderator-based explanations of heterogeneity. The findings contribute to the literature on sovereign rating determinants and have practical implications for fiscal monitoring, suggesting that debt management aimed at improving near-term servicing capacity matters for credit assessments in ways that are not fully captured by stock-based fiscal anchors. Full article

Firefighter training requires accurate posture monitoring to reduce injuries and improve performance assessment, yet traditional tracking methods suffer from high occlusion rates and the uniform appearance of trainees. To address these challenges, we propose an improved multi-target tracking algorithm that integrates YOLOX for detection, BlazePose for posture estimation, and a pose-constrained extension of DeepSORT. First, posture features are introduced into the association metric through a posture-cosine distance, which enhances discrimination between visually similar firefighters. Second, a pose-guided bounding-box correction is applied to ensure complete coverage of the human body region, improving the quality of extracted posture information. Experiments were conducted on a custom firefighter training dataset comprising 6602 labeled images and five multi-target video sequences (FM-1 to FM-5). The proposed method achieved a mean Average Precision (mAP) of 97.8% for detection and improved tracking performance compared to baseline DeepSORT, with MOTA rising from 74.72% to 82.96% and IDF1 from 74.77% to 82.36%. These results demonstrate that the algorithm effectively handles severe occlusion and appearance similarity, providing a reliable tool for posture tracking and behavior perception in firefighter training environments. Full article

Forests of Lualaba Province (DR Congo) form a compositionally complex mosaic of dry dense forest, gallery forest, and Miombo woodland. Yet, categorical land-cover maps impose discrete boundaries on these inherently continuous vegetation gradients, systematically discarding subpixel compositional information critical for forest monitoring and carbon accounting. The magnitude of this information loss at the landscape scale, however, remains largely unquantified. In this study, we train a Multi-Output Neural Network (MONN) using Sentinel-2 spectral and textural predictors (2025) to estimate the proportional cover of three forest types across the province. Model performance is benchmarked against a normalised Random Forest (RF) using spatial block cross-validation. Categorical information loss is quantified pixel-wise using two complementary metrics, dominant class proportion and Shannon compositional entropy, alongside a derived interpretive quantity, categorical information loss. The MONN slightly outperformed RF (R² = 0.648 vs. 0.630; RMSE = 0.224 vs. 0.229), yet the results reveal a fundamentally heterogeneous landscape structure. The mean dominant-class proportion was only 56.2%, indicating that categorical maps discard, on average, 43.8% of compositional information per pixel. Only 7.9% of forested pixels exceeded the 75% dominance threshold, while Shannon entropy reached 74.1% of its theoretical maximum, indicating that forest types coexist in near-equal proportions across most pixels. This renders categorical attribution structurally inadequate for most of the forested landscape. Across 92.1% of forested pixels, no single forest type achieved clear dominance. These results show that compositional mixing is the dominant structural condition of the landscape, and that compositional mapping is essential for representing tropical forest structure in heterogeneous drylands. By formally quantifying categorical information loss at the landscape scale, this study shows that continuous compositional mapping converts this structural ambiguity into a spatially explicit ecological signal, with direct implications for monitoring vegetation dynamics and biodiversity, suggesting a structural source of error in carbon stock estimation in tropical dry forests that warrants empirical validation. Full article

Rapid and non-destructive estimation of maize (Zea mays L.) leaf flavonoid (Flav) content is important for crop stress monitoring and precision agriculture. This study aimed to improve Flav estimation by integrating unmanned aerial vehicle (UAV)-based multispectral data, texture features, and phenological parameters across six key growth stages in the Guanzhong Plain, China. Maize Flav content was measured in situ using a Dualex Scientific+ meter, while canopy reflectance was acquired with a DJI M300 RTK UAV equipped with an MS600 Pro multispectral camera. A comprehensive feature set, including spectral bands, vegetation indices, texture features, texture indices, and logistic curve-derived phenological parameters, was constructed. Three feature selection methods, competitive adaptive reweighted sampling (CARS), the genetic algorithm (GA), and the successive projections algorithm (SPA), together with three regression models, partial least squares regression (PLSR), extreme gradient boosting (XGBoost), and convolutional neural network (CNN), were evaluated for Flav estimation. The results showed that integrating spectral, texture, and phenological information significantly improved model performance compared with spectral variables alone. CNN and XGBoost generally outperformed PLSR. Across the six growth stages, the stage-specific optimal models achieved coefficient of determination (R²) values ranging from 0.7749 to 0.8686 and residual prediction deviation (RPD) values ranging from 2.0046 to 2.6019, indicating high to outstanding predictive ability. The highest accuracy was obtained at R3 using the CARS-X_II-CNN model, with R² = 0.8686, root mean square error of validation (RMSEV) = 0.0382, and RPD = 2.6019. Texture features and phenological metrics, especially the start of season derived from the normalized difference vegetation index (NDVI_SOS) and the rate of senescence derived from the enhanced vegetation index (EVI_ROS), contributed substantially to model accuracy. In addition, maize Flav showed a unimodal response to nitrogen supply, with moderate nitrogen levels associated with higher Flav content. This study demonstrates the potential of UAV-based multisource feature integration and machine learning for accurate maize Flav estimation, and provides a useful framework for digital crop phenotyping and stress diagnosis. Full article

Background: Intracranial pressure (ICP) monitoring remains the cornerstone of neurocritical care in severe traumatic brain injury (TBI), yet its prognostic value as a standalone metric is limited. The Pressure Reactivity Index (PRx), a continuous measure of cerebrovascular reactivity derived from ICP and arterial blood pressure, may offer additional or complementary prognostic information. This systematic review aimed to compare the prognostic performance of PRx-derived metrics versus standard ICP monitoring for mortality and functional outcome in patients with TBI. Methods: A systematic search of PubMed, Web of Science, and Scopus was conducted for studies published between January 2000 and December 2025. Studies were eligible if they included adult TBI patients with continuous multimodal monitoring and reported comparative prognostic data for PRx- and ICP-based metrics. Risk of bias within the studies was appraised via the QUIPS tool, and the GRADE system was used to rate the strength of the evidence. Due to methodological heterogeneity, findings were synthesized narratively. Results: Nine studies were included. Applying a maximum-cohort estimation to account for overlapping registries, the pooled sample comprised a minimum of 1240 unique patients. In the majority of included studies, direct within-cohort head-to-head comparisons demonstrated that specific PRx-derived metrics—such as the individualized ICP threshold (iICP), Longest Continuous Duration of Autoregulatory Impairment (LCAI), Lower Limit of Reactivity (LLR), and time-integrated burdens (%Time > Threshold)—yielded stronger prognostic discrimination compared to standard ICP thresholds for both mortality (PRx: AUC 0.747–0.648 and ICP: AUC 0.660–0.614) and functional outcome. When added to established predictive models, PRx-derived metrics provided clinically meaningful incremental improvements in prognostic accuracy, with descriptive incremental AUC gains ranging from +0.039 to +0.170 across the six studies reporting model augmentation. Due to heterogeneity in baseline models, PRx-derived metrics, and patient populations, these findings are presented strictly as a descriptive range. Conclusions: PRx and PRx-derived cerebrovascular reactivity metrics-namely iICP, LCAI, LLR, and time-integrated burdens of autoregulatory failure—show potential to offer additive prognostic value beyond standard ICP monitoring in severe TBI. However, because current evidence is strictly observational and likely influenced by institutional confounders, it cannot currently support definitive clinical recommendations. Further prospective, multicenter studies utilizing standardized thresholds are necessary to confirm these associative findings and isolate their true prognostic value. Full article

Against the backdrop of global energy transition and the widespread adoption of Hydrogen-Blended Natural Gas (HBNG), the safety of urban gas pipeline networks faces severe challenges. This paper systematically reviews the research progress of numerical simulation in the field of natural gas pipeline safety, focusing on its core supporting roles throughout the “Leakage–Dispersion–Evolution–Consequence” disaster chain. First, it analyzes the kinetic modeling of high-pressure leakage holes and property corrections based on real gas equations of state, elaborating on the numerical characterization of HBNG multi-component transport. Second, it compares the dispersion mechanisms and environmental coupling modeling methods in typical scenarios such as buried porous media, confined spaces in utility tunnels, underwater environments, and urban building clusters. Third, it reviews leakage monitoring technologies based on physical field simulation and data-driven approaches (e.g., Convolutional Neural Network, Long Short-Term Memory), emphasizing the value of numerical simulation in constructing digital twin training sets. Furthermore, it explores the dynamic evolution of explosion flame–shock wave interactions and the evaluation models for secondary disaster consequences. Finally, the current research status of grid-based risk pre-warning and emergency response strategies is summarized. In conclusion, numerical simulation is not only a robust method for precisely quantifying and characterizing complex physical mechanisms but also a critical technological foundation for building smart and resilient energy cities. Future research should focus on the deep coupling of multi-physics fields, physics-informed learning, and the development of system-level integrated defense systems. Full article

This study presents a gate-to-gate life cycle assessment (LCA) of an industrial sweet cherry sorting and packing facility in Greece, directly addressing environmental sustainability in agri-food supply chains through data-driven impact quantification and improvement pathways in post-harvest operations. The assessment focuses on a gate-to-gate system boundary encompassing all processes inside the cherry sorting and packing facility, while upstream cherry production and downstream waste management are modeled and reported separately to provide system-level context. Core-stage hotspots are then analyzed in detail in the Results section, highlighting the dominant role of electricity use compared with packaging materials. The functional unit is defined as 1 kg of packed, market-ready cherries at the factory gate. Primary data are obtained from high-resolution, batch-level measurements of mass flows, energy use, water consumption, packaging materials and waste streams over a full processing season, structured as virtual sensor outputs. These sensor-informed operational data are combined with secondary life cycle inventory information from established databases to quantify climate change impacts and identify environmental hotspots across materials, energy, water, and waste, thereby delivering a quantified picture of environmental performance in the post-harvest stage. The results show that corrugated cardboard and associated packaging components are among the main contributors within the facility-level, gate-to-gate system, while the Core stage accounts for 28.43% of total GWP100. Upstream cherry production dominates the overall Upstream–Core–Downstream climate footprint with 70.61% of total impacts. Moreover, practical mitigation scenarios are modeled, including packaging optimization, partial substitution of grid electricity with photovoltaic generation, and increased water recirculation. Ιn the combined mitigation scenario, where packaging optimization, low-carbon electricity and improved water management are implemented simultaneously, total GWP100 decreases from 114,207.32 to 92,500.27 kg CO₂-eq (−19.0%) relative to the baseline, providing actionable sustainability improvements for industry stakeholders and supporting Sustainable Development Goals (SDGs) related to climate action and resource efficiency. In addition, the proposed virtual sensor architecture and data workflow support continuous monitoring, eco-efficiency management and near-real-time LCA implementation in post-harvest agri-food systems, enabling operational sustainability. Full article

The positioning and clamping process in aircraft assembly exhibits pronounced long-term temporal correlations and intense human–machine interactions. Consequently, assembly quality depends heavily on operator compliance and consistency. Capturing long-term, strongly correlated features in complex industrial environments remains a significant challenge. To overcome this, this study proposes a Long-Term Strongly Associated Action Recognition Network (LTSA-Net) tailored for aircraft assembly positioning and clamping tasks. Based on the C3D backbone, the model first incorporates the SimAM attention mechanism and BN modules to significantly enhance focus on critical spatiotemporal features. To address the challenge of capturing long-term temporal dependencies, LTSFEM is designed to extract global temporal information accurately. Furthermore, to balance structural lightweight design with real-time inference requirements, the CWSTB module is integrated to achieve substantial parameter compression. In addition, a dedicated aircraft assembly positioning and clamping dataset was constructed, and a robust training framework was established using the AdamW optimizer and Mixup data augmentation. Experimental results demonstrate that LTSA-Net achieves a recognition accuracy of 98.82% on the LTSA-Dataset, with a per-frame inference time of 42 ms, successfully meeting the dual requirements of high precision and real-time performance in industrial scenarios, and providing a practical technical solution for intelligent monitoring of aircraft assembly processes. Full article

Current landslide displacement prediction models often suffer from insufficient integration between physical mechanisms and data-driven approaches, weak model generalizability, and limited operational applicability. To address these issues, this study develops a physics-informed online learning framework for landslide displacement prediction. The core of this framework is a Physics-informed Long Short-Term Memory network (Phys-LSTM). By embedding discretized forms of the stress balance, creep constitutive, and kinematic equations as hard constraints into the LSTM’s gating mechanisms and loss function, the model ensures physically consistent predictions and enhanced interpretability throughout the learning process. Leveraging real-time data streams from the Sichuan Provincial Geological Hazard Monitoring and Warning Platform, we developed an online processing pipeline for real-time multi-source data ingestion, automated quality control, spatiotemporal alignment, and physics-informed feature engineering. A progressive three-stage learning algorithm was designed to support model cold-start, incremental training, and rolling prediction. Validation across 45 model-development landslide sites and one independent application case demonstrated the framework’s significant superiority over traditional models in displacement prediction accuracy (RMSE ≤ 1.78 mm, R² ≥ 0.96), cross-site generalization stability, and its capability to capture accelerated deformation phases. This research indicates that deeply integrating geomechanical prior knowledge into an online learning framework can effectively improve the reliability, interpretability, and operational applicability of landslide displacement prediction models, thereby providing methodological support for subsequent landslide early warning applications. Full article

Milk titratable acidity is a key indicator of raw milk freshness and quality, but its variation across different dairy animal species remains incompletely characterized. Based on 16,984 raw milk samples from eight dairy animal species (Holstein cow, goat, buffalo, camel, sheep, yak, donkey, and horse) collected within a retrospective raw milk quality monitoring framework in China from 2016 to 2024, this study provides a large-scale descriptive comparison of milk titratable acidity across species. Distinct titratable acidity profiles were observed among species, with camel and yak milk showing relatively high values, sheep, Holstein, and buffalo milk exhibiting intermediate values, and donkey and horse milk presenting markedly low values. Calendar-season-associated patterns also differed among species. Correlations between titratable acidity and milk components varied by species, with relatively stronger positive associations with protein and solids-not-fat (SNF) in several ruminant milks, suggesting that milk composition may contribute to differences in titratable acidity. However, because this study was based on an unbalanced observational dataset with limited animal-level, farm-level, feeding, management, physiological, and environmental metadata, these observations should be interpreted as descriptive and exploratory patterns rather than causal biological mechanisms. This dataset provides preliminary reference information for future studies on species-associated variation in raw milk titratable acidity and for discussions on species-specific raw milk quality evaluation. Full article

Lubricating oil monitoring provides continuous health information for early fault warning and maintenance decision-making in industrial gas turbines. However, real-world multi-sensor monitoring streams exhibit pronounced nonstationary thermodynamic drifts that often obscure subtle high-frequency residuals containing critical incipient degradation signatures. Prevailing data-driven monitoring models typically struggle to separate these macroscopic trends from stochastic wear-related fluctuations, and their restrictive distributional assumptions are often inadequate for the heteroscedastic and heavy-tailed nature of industrial residuals. To address these challenges, this study proposes ResAD-Net, a framework for early fault warning in nonstationary multi-sensor oil monitoring that combines trend–residual decoupling, trend-conditioned residual modeling, and residual-domain dependency learning. Specifically, a signal trend–residual decoupling strategy is adopted to separate slowly varying operational trends from stochastic residual fluctuations captured by the sensors, thereby exposing residual information that is more sensitive to incipient degradation. On this basis, a trend-conditioned diffusion model is introduced to characterize state-dependent, skewed residual distributions and generate residual sample ensembles for nonstationary monitoring. Meanwhile, a graph-based variational autoencoder is employed to learn latent intersensor dependency structures from the residual domain, providing diagnostic cues for temporal risk evolution analysis and sensor-level inspection. Experiments on a real-world industrial oil-monitoring record show that the proposed framework achieves an average F1-score of 0.985 with no observed false positives in the predefined pre-alarm reference interval of the finite test set. In addition to accurate anomaly detection, ResAD-Net captures early residual distributional shifts before clear macroscopic deviations emerge and provides diagnostic association cues for interpreting oil-monitoring changes around the system-level alarm. Full article

Accurate bathymetric information for inland lakes is essential for water resource management, ecological monitoring, and environmental research. However, the accuracy and robustness of remote sensing-based bathymetric retrieval are often constrained by the complex optical properties of inland waters and the limited representation of conventional inversion features. To address these challenges, this study systematically compared the performance of a multiband logarithmic ratio model and three machine learning models, including Random Forest (RF), XGBoost, and AdaBoost, for inland lake bathymetric retrieval. Furthermore, a synergistic retrieval framework integrating chlorophyll-a concentration (Chla) and a Water Optical Index (WOI) was proposed. The results show that: (1) The overall accuracy of the Random Forest, XGBoost, and AdaBoost models constructed with the integration of chlorophyll-a concentration and WOI ($R^2=0.93$, $0.93$, and $0.91$; MAE $=0.06$ m, $0.07$ m, and $0.12$ m; RMSE $=0.14$ m, $0.14$ m, and $0.16$ m) outperforms that of models using only multispectral band information ($R^2=0.93$, $0.91$, and $0.82$; MAE $=0.06$ m, $0.07$ m, and $0.14$ m; RMSE $=0.14$ m, $0.16$ m, and $0.22$ m). Moreover, all these machine learning models significantly outperform the traditional numerical model ($R^2=0.27$; MAE $=0.29$ m; RMSE $=0.45$ m), with the Random Forest model achieving the best overall performance. This indicates that the proposed method offers higher applicability and retrieval accuracy in complex inland lake environments. (2) The optimal Random Forest model integrating chlorophyll-a concentration and WOI achieved high-precision bathymetric inversion for inland lakes ($R^2=0.93$, MAE $=0.06$ m, RMSE $=0.14$ m). Based on the three-dimensional bathymetry derived from this model, the estimated lake storage capacity was $1072.11\times {10}^4$ m${}^3$, compared with a measured volume of $1094.27\times {10}^4$ m${}^3$, yielding a relative error of $-2.03\%$. This result provides reliable and highly accurate data to support water resource management. Full article

Multi-modal land cover classification plays an important role in remote sensing applications such as urban monitoring and environmental analysis. By integrating complementary information from hyperspectral imagery (HSI) and LiDAR data, multimodal learning can significantly improve classification performance. However, existing Transformer-based fusion methods often suffer from high computational complexity and inefficient cross-modal interaction modeling, which limits their applicability in resource-constrained scenarios. To address these challenges, we propose LMFusion, an efficient framework for multimodal feature learning. Specifically, LMFusion enables efficient bidirectional feature interaction through a linear-complexity cross-attention mechanism and enhances long-range spatial-spectral representation learning with Mamba-based state space modeling, thereby achieving effective multimodal dependency modeling with linear computational complexity. In addition, a selective quantization-aware optimization strategy is introduced to support multiple bit-width settings (down to 1-bit), yielding a more compact and efficient model while improving representation robustness under low-bit constraints. Extensive experiments on the Houston2013, MUUFL, and Augsburg datasets demonstrate the effectiveness of LMFusion. It achieves overall accuracies of 95.84%, 94.95%, and 99.05%, respectively, consistently outperforming representative multimodal classification methods and showing strong potential for accurate and efficient multimodal remote sensing classification. Full article