Search Results (21,819)

In natural scenes, the high resemblance between unripe tomatoes and foliage, combined with severe occlusion, challenges current semantic segmentation models, causing poor accuracy and indistinct boundaries. To overcome this, we propose CMAE-UNet, a camouflage suppression and edge enhancement model based on the UNet architecture. Specifically, a Global-Local Integrated Spatial Attention (GLISA) encoder merges dual-branch dilated convolutions, residual structures, and an Efficient Multi-scale Attention mechanism to expand receptive fields and highlight targets in complex backgrounds. Furthermore, a Frequency-Domain Feature Enhancement (FFE) module leverages the Fast Fourier Transform to separate and adaptively enhance distinct frequency components, effectively mitigating camouflage interference. Additionally, a Directional Edge Enhancement (DEE) module uses three-directional learnable convolutions and spatial attention to sharpen indistinct target contours. Evaluated on a custom Tomato dataset encompassing five complex scenarios, CMAE-UNet outperforms 12 prominent methods in mIoU, Dice, and Sen metrics, yielding smoother and more precise segmentation boundaries. The model robustly withstands field interference, providing strong technological support for automated tomato detection, intelligent harvesting, and growth monitoring. Full article

Accurate and timely monitoring of terrestrial oil spills is vital for ecological conservation and safe oilfield operations. To address the challenges of segmenting terrestrial oil spills in UAV remote sensing imagery, including blurred boundaries, irregular shapes, and complex background interference, we propose Fluid-SegFormer, a fluid-aware semantic segmentation model based on the lightweight SegFormer architecture. Fluid-SegFormer employs a Mix Transformer (MiT-B0) encoder to extract hierarchical multi-scale features and integrates a hierarchical fluid-aware optimization framework. Specifically, the Local Noise Gating (LNG) module suppresses background noise, the Horizontal–Vertical Perception Attention (HVPA) module enhances the structural representation of irregular oil spill regions, and the Fluid Soft Boundary Refinement Decoder (FSBRD) recovers fine boundary details. Experiments on a newly constructed high-resolution UAV terrestrial oil spill dataset demonstrate that Fluid-SegFormer achieves an mIoU of 87.84%, an IoU of 77.56%, and a Precision of 91.42%, effectively balancing computational efficiency and segmentation accuracy. These results demonstrate the potential of Fluid-SegFormer for practical deployment in UAV-based oil spill monitoring on edge devices. Full article

This study proposes a surrogate model, Graph-Structured Physics-Informed Deep Operator Network (GS-PI-DeepONet), to simulate two-dimensional hydrodynamics in a tidal river network. The model was coupled with Graph Convolutional Networks (GCNs) for spatial feature extraction and Long Short-Term Memory (LSTM) network for temporal prediction, with 2D shallow-water equations (2D SWEs) embedded as physical constraints. To handle complex river network topologies, a mapping mechanism was proposed to transform discrete irregular boundaries into differentiable neural network constraints. A dynamic weighting strategy was developed to improve model training efficiency. GS-PI-DeepONet was applied to a river network within the Pearl River Basin in Zhuhai. Trained on high-fidelity Delft3D data, it achieved precise flow field reconstruction and millisecond-level extrapolation predictions, outperforming traditional data-driven models. The model can be a valuable tool for real-time hydrodynamic simulations and flood management strategies in tidal river networks. Full article

With the widespread adoption of cloud computing technology, modern cloud platforms have become increasingly complex and dynamic, posing significant challenges for efficient resource management. Accurate forecasting of cloud resource loads has therefore become essential for improving service quality, optimizing resource utilization, and reducing operational costs. To address the intrinsic characteristics of cloud load time series, including nonlinear fluctuations, multi-scale temporal dependencies, and redundant high-dimensional features, this paper proposes the MST-iTransformer model, which integrates multi-scale temporal encoding, sparse attention, and adaptive feature selection mechanisms. Specifically, a multi-scale temporal encoding module is developed to capture and fuse temporal dependencies across multiple periodic scales. Furthermore, an adaptive feature selection module is introduced to dynamically assign importance weights to resource features, enhancing informative variables while suppressing redundant ones. Meanwhile, a sparse attention mechanism is incorporated to reduce computational overhead while maintaining forecasting accuracy. The proposed model is evaluated on the Alibaba Cluster Trace dataset. Experimental results demonstrate that MST-iTransformer achieves MSE, RMSE, and MAE values of 0.5559, 0.7456, and 0.4968, respectively. Compared with the original iTransformer, the proposed model achieves simultaneous reductions in prediction errors and inference latency, validating the effectiveness of the multi-scale temporal encoding, sparse attention mechanism, and adaptive feature selection modules in improving forecasting accuracy and computational efficiency. These improvements provide reliable prediction support for resource scheduling and elastic scaling in cloud data centers. Full article

Pesticides pose significant hazards to aquatic ecosystems and public health; therefore, understanding their fate in aquatic systems is critically important. Photochemical processes driven by direct and indirect photolysis, particularly hydroxyl radical (^•OH) reactions, play a pivotal role in the transformation of these contaminants in natural waters. This study employed an efficient and selective ^•OH production technique using a high-power UV light-emitting diode (UV-LED) combined with nitrite photolysis to determine the second-order reaction rate constants between ^•OH and selected pesticides (k_X,OH). This approach enabled reliable estimation of the indirect photodegradation rate constants (k_IP) for the selected pesticides in aquatic systems. In addition, direct photodegradation rate constants (k_DP) of the selected pesticides were determined under simulated sunlight conditions using a solar simulator equipped with a 500 W xenon lamp. The photochemical half-lives of selected pesticides in river water and seawater were calculated from k_DP and kIP under assumed steady-state HO^• concentrations. The results demonstrated that direct photolysis rate constants of the selected pesticides ranged from 1.62 × 10⁻⁷ to 5.52 × 10⁻⁴ s⁻¹. The second-order reaction rate constants between the investigated pesticides and ^•OH ranged from 0.045 × 10⁹ to 14.8 × 10⁹ M⁻¹ s⁻¹. Estimated half-lives under direct photolysis in seawater ranged from hours to days, whereas half-lives attributed to indirect photolysis in seawater extended to several years. In contrast, half-lives of pesticides in river water ranged from hours to days for indirect photolysis. Under the assumed steady-state ^•OH concentrations, direct photolysis generally produced shorter calculated half-lives than the ^•OH pathway in seawater, whereas the higher assumed ^•OH concentration substantially reduced the calculated indirect-photolysis half-lives in river water. These findings should be interpreted as condition-specific kinetic comparisons rather than direct measurements of environmental persistence. Full article

Energy efficiency (EE) is widely promoted as a climate-mitigation strategy, yet improvements in energy intensity emerge from interactions among institutions, technologies, economic structures, and environmental conditions and cannot be attributed to efficiency policies alone. This study examines the evolution of EE governance in Chile between 1992 and 2022, treating national energy intensity as an emergent property of a path-dependent sociotechnical system. A two-component mixed-methods approach combines a historical-institutional reconstruction of governance with a parsimonious econometric analysis of short-run changes in energy intensity, integrated through a causal loop diagram. The historical analysis shows EE governance evolving through path dependence, critical junctures, layering, and policy feedback, strengthening state coordination without transforming the underlying market-oriented regime. Across four specifications, only carbon intensity is robustly associated with short-run changes in energy intensity; structural, extractive, and sociotechnical variables show no robust association. A comparison of the two strands reveals an asymmetry: the institutional trajectory coincided with aggregate change in the carbon–energy dimension, while no comparable short-run signature was detected along the direct demand-side routes examined. These findings reframe energy intensity as a system-level property and indicate that efficiency policies should be complemented by demand-side governance, structural change, and sufficiency to support sustainability transitions. Full article

Digital infrastructure (DI) plays a dual role in the low-carbon transition. It supports economic operation but also consumes substantial energy. This study explores DI’s impact on carbon emission efficiency (CEE) using data on 41 cities in China’s Yangtze River Delta from 2011 to 2024. The methods used in this study include a two-way fixed effects model, mediation analysis, a panel threshold model, and a spatial Durbin model. The results show that the impact of DI on CEE is U-shaped. Industrial upgrading and technological innovation are the potential channels through which DI affects CEE. Energy efficiency has a single threshold value of 8.533. DI enhances CEE when energy efficiency exceeds this threshold. Spatial analysis indicates that both the direct and indirect effects of DI follow a U-shaped pattern. Heterogeneity analysis indicates that the environmental impact of DI varies depending on resource endowments, policy environments, and economic development levels. This study provides insights for global urban agglomerations to balance digital transformation and sustainable development. Full article

Artificial intelligence (AI) has generated considerable excitement in radiology, with claims of transformative improvements in diagnostic accuracy, workflow efficiency, and clinical decision support. However, a critical gap persists between AI’s theoretical promise and its real-world performance, particularly in complex, high-stakes scenarios such as cancer-associated thromboembolism (CAT). CAT is a leading cause of morbidity and mortality in oncology patients, yet it remains underdiagnosed on routine imaging. This paper provides a general radiology critique of current AI applications, then narrows focus to CAT management. This review additionally evaluates AI’s role in incidental pulmonary embolism detection, risk stratification, and treatment decision support. While AI demonstrates sensitivity gains, it faces substantial limitations: data heterogeneity, lack of prospective validation, poor generalizability across cancer subtypes, and integration challenges with clinical workflows. Therefore, AI is not yet a reliable standalone tool for CAT management, but may serve as an adjunct if clinically validated, explainable, and embedded within multidisciplinary frameworks. Full article

Water contamination with surface-active agents, such as sodium lauryl sulfate (SLS), represents a serious environmental concern, driving the need for efficient and low-cost alternative adsorbents as a step toward sustainable waste valorization. In this study, waste biomass derived from plum stones, date stones, and walnut shells was successfully transformed into activated carbons via chemical activation using potassium hydroxide (KOH) at 850 °C with a 1:1 impregnation ratio. The synthesized materials underwent comprehensive physicochemical characterization utilizing TG-DSC, elemental analysis, Boehm titration, SEM, TEM, and nitrogen physisorption (BET), whereas their adsorption performance was evaluated against aqueous SLS solutions across various concentrations. The obtained results reveal a predominantly microporous structure with a high specific surface area, reaching up to 1059.01 m²/g for ACdate. The equilibrium adsorption data were well described by the Langmuir isotherm model, which yielded model-estimated asymptotic adsorption capacities (q_m) of 219.70 mg/g for ACwalnut, 178.25 mg/g for ACdate, and 57.80 mg/g for ACplum. These values represent Langmuir-derived model parameters rather than experimentally attained adsorption capacities within the investigated concentration range. Notably, despite having a lower specific surface area than ACdate, ACwalnut exhibited the highest Langmuir-estimated q_m, which may be associated with its structural balance and well-developed mesoporous network (0.210 cm³/g), facilitating the intraparticle transport of SLS molecules. These findings highlight that high efficiency originates from a synergistic combination of accessible porosity, a mesoporous transport network, hydrophobic character, and specific surface functional groups, demonstrating the exceptional potential of these activated carbons for anionic surfactant wastewater remediation. Full article

Fast charging of lithium-ion batteries requires balancing charging efficiency with electrochemical safety to minimize degradation and lithium plating. Conventional charging strategies and existing reinforcement learning approaches often lack physical consistency and model interpretability, limiting their applicability in safety-critical battery management systems. This paper proposes a physics-informed Attention-Proximal Policy Optimization (Attention-PPO) framework for intelligent battery fast charging by integrating the Single Particle Model with Electrolyte (SPMe) with a transformer-based attention mechanism. SPMe provides physically meaningful battery state transitions, while the attention-enhanced PPO dynamically learns informative electrochemical representations for charging control. To improve transparency, a monotonic XGBoost surrogate model is employed for lithium-plating risk estimation, and the learned policy is further distilled into an interpretable decision tree. Experimental results demonstrate that the proposed Attention-PPO achieves substantially faster and more stable policy convergence than the baseline PPO, reaching convergence at episode 37 compared with episode 82 for PPO, corresponding to a 54.9% reduction in training episodes. The reward standard deviation is also reduced from 0.132 to 0.041, indicating 68.9% lower reward variability. In terms of electrochemical safety, Attention-PPO achieves a mean plating overpotential of +0.023 V compared with −0.015 V for PPO, providing a 38 mV improvement and a positive safety margin against lithium plating. Compared with conventional CC-CV and CC-COP controllers, the proposed framework requires a longer charging duration because it prioritizes electrochemical safety; however, it consistently maintains positive plating overpotential while achieving reliable charging performance. Compared with conventional CC-CV, CC-COP, and standard PPO controllers, the proposed framework provides a larger electrochemical safety margin while preserving reliable charging performance. Transformer attention analysis and policy distillation provide interpretable representations of the learned charging policy, while SHAP analysis characterizes feature contributions within the auxiliary plating-risk estimator. The proposed framework provides an effective and explainable physics-informed reinforcement learning solution for safe lithium-ion battery fast charging, offering a promising approach for next-generation intelligent battery management systems. Full article

The efficient development of heavy oil reservoirs is challenged by the high viscosity and poor mobility of heavy oil. Although multicomponent thermal fluid technologies involving heat, gas, and chemical agents have demonstrated potential advantages over conventional steam-based recovery methods, the microscopic synergistic mechanisms responsible for viscosity reduction remain insufficiently understood. Therefore, this study investigates the synergistic mechanisms by which heat, an alkane solvent (C₁₁H₂₄), and CO₂ reduce heavy-oil viscosity. Heavy oil from the Shengli Oilfield was selected as the research object, and rheological experiments were combined with molecular dynamics simulations to systematically analyze viscosity variations and their underlying microscopic mechanisms under different conditions. The experimental results demonstrate that increasing temperature significantly reduces heavy oil viscosity, and a characteristic transition in viscosity reduction behavior occurs at approximately 100 °C. At 90 °C, the addition 5 wt% oil-soluble viscosity reducer C₁₁H₂₄ decreases the heavy oil viscosity to 442.2 mPa·s, corresponding to a reduction rate of 83%. The solubility of CO₂ increases markedly with pressure, and at 30 MPa, the viscosity reduction exceeds 99%. The combined effects of these three factors exhibit superior viscosity-reduction performance. Molecular dynamics simulation results indicate that CO₂ and the viscosity reducer synergistically weaken the π-π stacking interactions of asphaltenes and resins in heavy oil, transforming heavy components from locally aggregated states into more uniformly dispersed configurations. Meanwhile, the intermolecular interaction energy and cohesive energy density decrease, indicating weakened molecular interactions and enhanced diffusion behavior. These results demonstrate that the synergistic viscosity-reduction mechanism of heat–gas–agent systems is mainly associated with structural disaggregation, interaction weakening, and diffusion enhancement. This study provides molecular-level insights into multicomponent thermal fluid-assisted heavy oil recovery and offers theoretical support for improving heavy oil development efficiency. Full article

Smart lighting is becoming a strategic component of intelligent and low-carbon urban infrastructure because it combines efficient illumination with connectivity, remote management, and continuous operational monitoring. This study proposes a reproducible unit-level methodological framework that transforms field records from a functional smart-lighting installation into traceable indicators of electrical performance, energy efficiency, avoided emissions, preliminary economic benefit, sensitivity, and conditional scalability. The approach treats the luminaire not only as an electrical load, but as a monitored urban energy node whose operation can be validated, characterized, and compared under planning-oriented control scenarios. The methodology integrates data preprocessing, electrical consistency assessment, representative baseline definition, scenario-based energy modeling, explicit environmental conversion, and conditional scaling to homogeneous lighting assets. The results reveal a stable electrical operating regime and show that managed operating conditions can generate sustained reductions in energy use and associated environmental impacts while preserving analytical transparency between measured variables and scenario-derived indicators. Sensitivity and multivariable analyses further support the robustness of the unit-level interpretation and highlight the value of monitored lighting data for comparative decision-making. The framework therefore provides a technically grounded reference for smart-city lighting management, energy planning, and scalable infrastructure assessment, with relevance to the objectives of SDG 7, SDG 11, and SDG 13. Overall, the study contributes an original data-driven perspective for integrating IoT-enabled lighting, remote supervision, and sustainability-oriented urban management within a common analytical structure. Full article

In recent years, interest in artificial intelligence has grown significantly, particularly in the development of advanced computational models for supply chain decision-making. Inventory management is one of the areas in which intelligent algorithms can support demand forecasting, replenishment, stock control, and operational optimization under uncertainty. This study presents an updated systematic literature review of intelligent algorithms applied to inventory management. The review followed PRISMA 2020 guidelines and combined database searches in Scopus, ScienceDirect, Web of Science, IEEE Xplore, SpringerLink, Taylor & Francis, and complementary manual searching. The original search covering January 2020 to December 2024 was updated in July 2026 to include studies published or available online up to June 2026. After applying strict eligibility criteria, 37 primary studies with quantitative evidence were included. The updated corpus confirms the predominance of deep learning, reinforcement learning, and hybrid intelligent models, while also showing the recent emergence of Transformer-based, graph neural network, multi-agent reinforcement learning, and prescriptive analytics approaches. The most frequent application areas were inventory control, inventory optimization, replenishment decision-making, and demand forecasting. Reported improvements were mainly associated with cost efficiency, service level, stockout reduction, and system performance; however, the magnitude of improvement varied across algorithms, data sources, sectors, and simulation or real-world settings. Overall, intelligent algorithms represent a relevant tool for improving inventory management, but their adoption requires careful validation, transparent reporting, and alignment with the operational context. Full article

Accurate identification of multiphase flow patterns in vertical pipelines is critical for operational safety and efficiency in the oil and gas industry. Yet, conventional methods struggle with subjectivity and transitional regimes. This study develops and integrates three neural network architectures: a convolutional neural network (CNN) for spatial features, a transformer neural network (TNN) for long-range dependencies, and a physics-encoding network (PEN) for embedding physical constraints. These are combined into a hybrid framework trained on an experimental dataset of 2131 images from a wire mesh sensor, annotated using a semi-automated pipeline. Results show the hybrid model achieved 95.91% test accuracy with a macro F1-score of 0.96, the highest of the four models evaluated, with its main advantage in transitional regimes. A multi-seed ablation shows that the convolutional branch provides the dominant discriminative signal, while the transformer and physics-inspired branches added complementary improvements that are consistent across runs. This hybridisation mitigates individual model weaknesses, with the physics-inspired branch acting as a spatial regulariser that improves interpretability, providing a robust and objective tool for reliable pipeline monitoring. Full article

Emerging industrial agglomeration drives green transformation through knowledge spillovers and economies of scale. However, its effects exhibit pronounced nonlinearity and heterogeneity, shaped by spatial externalities and development stages. This paper investigates 19 Chinese urban agglomerations over the period 2014 to 2023. Kernel density estimation based on enterprise-level Point-of-Interest (POI) data is used to characterize spatial agglomeration patterns across eight emerging sectors. A two-stage dynamic network super-efficiency SBM model decomposes Green Transition Efficiency (GTE) into resource utilization and pollution control sub-stages. An XGBoost-SHAP-GEO analytical framework, combined with partial dependence analysis, then identifies nonlinear driving mechanisms. The main findings are as follows: First, emerging industrial agglomeration intensifies and polarizes toward the eastern coast, whereas GTE displays a “high-west, low-east” pattern. This produces a significant spatial mismatch, rooted in the near-saturation of environmental carrying capacity in eastern regions, where congestion effects exceed knowledge spillover dividends. Second, geographic characteristics constitute the primary factor shaping GTE and operate through nonlinear interactions with industrial agglomeration and R&D investment. Notably, their moderation direction is reversible, suggesting that geographic endowments should be understood as “conditional assets” rather than fixed advantages. Third, nonlinear patterns across sectors are highly heterogeneous. The bio-industry is the only sector to achieve a J-shaped positive breakthrough. Information technology and new materials exhibit persistent inhibition, while related services display an extremely narrow threshold window with the deepest negative reversal. Thus, “moderate agglomeration” is a multidimensional concept that shifts dynamically with industry type and regional endowment. Fourth, driving mechanisms display stage-dependent evolution. The incubation stage relies on natural endowments and basic industrial pull, with the green bottleneck residing in resource utilization efficiency. The growth stage faces multiple tensions from coexisting positive and negative effects. The optimization stage shifts toward R&D innovation and industrial greening, marking a qualitative transformation from MAR externalities to Jacobs externalities. In addition, the non-significant linear coefficient in the 2SLS instrumental variable test is consistent with the inverted U-shaped nonlinear finding, further validating the necessity of a nonlinear analytical framework. These findings provide differentiated governance evidence for balancing industrial agglomeration with green sustainable development. Full article