Search Results (4,095)

During the fabrication of prestressed concrete beams, the quality and positional accuracy of the laid corrugated ducts (or prestressing ducts) directly influence the load-bearing capacity and durability of the beams. However, traditional manual inspection is inefficient, highly subjective, and difficult to achieve full coverage. To address this problem, this paper proposes an automated detection method that integrates improved YOLOv11-based pose estimation, robust curve fitting, and image stitching techniques. The method automatically identifies duct positions and evaluates laying quality. By incorporating the SE channel attention mechanism and the SPPFCSPC multi-scale pooling module, the SC-YOLOv11 model is developed, which significantly enhances the detection accuracy of slender corrugated pipe key points in environments with dense rebar occlusion. The RANSAC algorithm is employed to fit curves to the predicted key points, effectively suppressing the influence of outliers. Furthermore, the SIFT algorithm is used for precise stitching of drone-captured segmented images, which are then transformed into a unified front orthographic coordinate system of the entire box girder via perspective transformation, enabling accurate reconstruction of the corrected 2D layout of corrugated ducts across the full beam. Ablation experiments using 5-fold cross-validation demonstrate that SC-YOLOv11 improves mAP₅₀ and mAP_50–95 by 2.6% and 1.2%, respectively, with statistical significance (paired t-test, p < 0.01). The model achieves a per-image inference time of 6.37 ms, with 4.34 M parameters and 8.1 GFLOPs, meeting real-time requirements. In a 30 m prefabricated box girder field application, the measured section trajectory fitting curves of the corrugated ducts were compared with the design alignment, successfully identifying two abnormal locations where the laying deviation exceeded the allowable threshold. Cross-validation with on-site inspector records shows that over 92% of the measurement points agree within ±10 mm. This method achieves a fully automated analysis chain from key point detection and curve fitting to deviation quantification, providing an efficient, non-contact, and traceable intelligent tool for quality control of bridge prestressed systems. Full article

The construction industry faces significant challenges in materials management, including inefficient supply chains and limited adoption of Circular Economy (CE) goals, which blockchain may address through automated tracking and verification systems. This systematic review examines blockchain technology applications in construction materials management to support CE strategies. Following PRISMA guidelines, 138 articles were selected from 1891 publications across four databases covering 2018 to 2025. The findings present a lifecycle-based framework across five building stages integrating smart contracts, IoT sensors, digital material passports, and tokenized waste exchange systems. Blockchain enables automated supply chain transparency, eliminates manual verification, and facilitates continuous material tracking. This research contributes by transforming conventional materials management into autonomous, data-driven workflows through a blockchain-enabled framework that systematically maps automated solutions for tracking, compliance, and circular resource flows across five building lifecycle stages, enabling practitioners to implement automated CE strategies. Full article

Accurate crop type mapping in Sub-Saharan Africa is a challenging task, due to the presence of smallholder farming systems characterized by fragmented landscapes and heterogeneous cropping practices. Persistent cloud cover, particularly significant during the cropping season, systematically limits the exploitation of optical satellite image time series, making things even harder. This study proposes MSF-Net (Multimodal Sentinel Fusion Network), a convolutional neural network-based late-fusion framework that combines Sentinel-1 synthetic aperture radar and Sentinel-2 multispectral time series for multi-class crop classification in the complex agricultural landscapes of central and northern Benin. The model was evaluated across six sites and three growing seasons (2022–2024) covering 12 land cover classes and compared with a Sentinel-2-only Temporal Convolutional Neural Network (TempCNN), a SAR-only baseline (S1-Branch), an ablated version of the proposed method, and two external state-of-the-art multimodal architectures, TSViT and TWINNS. MSF-Net achieved the highest or joint-highest overall accuracy in 10 of 14 site–year configurations, with overall accuracy ranging from 82.61% to 91.15% and kappa coefficients from 0.79 to 0.89, consistently outperforming both external baselines across all site–year configurations. The largest gains over TempCNN reached up to 30 percentage points for spectrally ambiguous classes such as Shrubby Savannah, Cotton, and Open Forest. In addition, MSF-Net produced more spatially coherent maps, with reduced salt-and-pepper noise, improved parcel-level homogeneity, and fewer modality-specific artefacts. These results demonstrate the value of SAR-optical fusion for operational crop monitoring in tropical West Africa. Full article

Sub-Mediterranean ecosystems along the northeastern Black Sea coast remain insufficiently studied despite their importance for the regional carbon cycle and their sensitivity to human disturbance. This study presents an exploratory case study of aboveground biomass carbon stocks in two contrasting plant communities at the carbon-monitoring supersite of the Southern Branch of the Shirshov Institute of Oceanology RAS in Gelendzhik, Russia. The aboveground observations were interpreted together with soil carbon stocks and greenhouse gas fluxes previously reported for the same permanent plots. The investigated communities comprised a terraced pine–oak plantation (SP1) and a natural pubescent-oak shiblyak (SP2). Vegetation was surveyed using the Braun-Blanquet approach. Tree biomass was estimated using the generalized allometric equation of Chave et al., while aboveground biomass was converted to carbon using carbon concentrations determined by elemental analysis of sampled plant tissues. Herbaceous vegetation and forest-floor material were quantified by elemental analysis. Within-plot spatial heterogeneity was visualized using Python-based radial-basis-function (RBF) interpolation for exploratory mapping of local carbon distribution. Because each vegetation type was represented by a single 400 m² plot, all comparisons are site-specific and exploratory and should not be extrapolated beyond the investigated communities. The aboveground carbon density was 64.1 t C ha⁻¹ in the terraced pine–oak stand and 16.1 t C ha⁻¹ in the natural shiblyak. Woody biomass dominated the plantation, whereas litter and herbaceous vegetation contributed proportionally more to the carbon pool in the shiblyak. Previously published soil organic carbon stocks at the terraced site exceeded the measured aboveground carbon pool, whereas the natural rendzinas were characterized by substantial carbonate-associated inorganic carbon stocks. C/N ratios of organic substrates ranged from 11 to 100, indicating a pronounced variation in substrate quality: higher C/N values in woody and coniferous fractions were consistent with slower decomposition, whereas lower C/N values in herbaceous material were consistent with faster mineralization. Exploratory spatial mapping revealed within-plot variation in aboveground carbon of up to two orders of magnitude. Together, these site-specific observations provide the first integrated description of aboveground biomass carbon together with previously reported soil carbon pools for contrasting sub-Mediterranean vegetation on the northeastern Black Sea coast. Despite its exploratory nature, this study establishes a baseline for future replicated, spatially explicit investigations of ecosystem carbon dynamics in sub-Mediterranean ecosystems along the northeastern Black Sea coast. Full article

This paper presents the design and preliminary field validation of a portable, low-cost Internet of Things (IoT) station for georeferenced soil nutrient profiling in agricultural environments. The proposed system integrates a digital RS-485 NPK soil sensor, an ESP32 microcontroller, and a SIM7000G GSM/GPS module to enable on-site acquisition and real-time transmission of nitrogen (N), phosphorus (P), and potassium (K) measurements using the MQTT protocol. Data are serialized in JSON format and transmitted to a ThingsBoard cloud platform for remote storage and visualization. The portable architecture supports manual spatial sampling across multiple locations without reliance on fixed infrastructure, making it suitable for small- and medium-scale agricultural contexts with limited connectivity. Preliminary testing in a controlled lemon plantation demonstrated stable GSM connectivity, successful geotagging, and consistent cloud-based visualization, with an average acquisition–transmission cycle of 30–45 s per measurement. Spatial heat maps generated from collected data illustrate the system’s capability for indicative nutrient mapping. Although laboratory-grade validation is ongoing, the results confirm the technical feasibility of integrating low-cost sensing, cellular communication, and georeferenced data acquisition into a compact IoT unit. The system establishes a foundation for future calibration, large-scale field validation, and decision-support applications in precision agriculture. Full article

Taking Lankao County as a case study, this study extends the established StrawFeed framework by linking NHPI-derived harvest dates to spatially differentiated straw release and seasonal equipment scheduling. Time-series Sentinel-2 observations were used to characterize harvest timing, while observed road routes were used to calibrate the detour parameter. The mapped harvest progression was compared with reported aggregate harvest progress and provided spatially differentiated timing inputs for logistics simulation. The results demonstrate that: (1) The spatial-economic framework effectively refine the spatial representation of traditional models in complex fragmented plots. (2) Under the annual shared-fleet baseline, the simulated delivered cost was 28.40 USD t⁻¹. (3) Route calibration yielded a road-detour coefficient of 1.36, demonstrating the effect of road geometry on transport distance. The simulated cost represents the result under the specified demand, road, equipment, and price assumptions rather than a causal penalty attributable to land fragmentation. Consequently, we suggest that land management policies should shift toward “precision-targeted” subsidies, and rural spatial planning should prioritize the layout of decentralized pre-processing hubs to optimize spatial logistics efficiency and economic viability. Full article

Background/Objectives: This study aimed to develop and internally validate multiparametric MRI radiomics models for predicting volumetric progression following Gamma Knife radiosurgery (GKRS) in sellar–parasellar meningiomas and to evaluate the incremental value of diffusion-derived features beyond contrast-enhanced imaging. Methods: Fifty-four patients underwent pretreatment and approximately 6-month post-treatment MRI, including contrast-enhanced T1-weighted imaging (T1C-WI) and apparent diffusion coefficient (ADC) maps. Whole-tumor segmentations were reviewed by two neuroradiologists by consensus. Radiomic features were extracted using PyRadiomics, and delta features were calculated as post-treatment minus pretreatment values. Elastic-net logistic regression models were evaluated using repeated nested cross-validation with five-fold inner and outer loops repeated 10 times. The primary endpoint was volumetric progression, defined as a >20% volume increase at 3 years. Results: At 3 years, 9 tumors (16.7%) progressed, 25 (46.3%) remained stable, and 20 (37.0%) regressed. The ΔT1C-WI model showed the highest repeated nested cross-validation performance, with a mean AUC of 0.861 ± 0.065, an accuracy of 0.846 ± 0.037, and an F1 score of 0.579 ± 0.097. Averaged patient-level out-of-fold predictions yielded an AUC of 0.914 (95% CI, 0.822–0.980), a sensitivity of 77.8%, and a specificity of 88.9%. The ΔADC model showed moderate discrimination, whereas combining ΔT1C-WI and ΔADC features did not improve performance. Conclusions: Delta radiomics derived from 6-month post-treatment T1C-WI may help identify sellar–parasellar meningiomas at risk of 3-year volumetric progression after GKRS. These findings suggest that 6-month ΔT1C-WI radiomics may support early risk stratification, but its clinical value requires external validation and prospective evaluation. Full article

Background/Objectives: Invasive intracranial pressure (ICP) monitoring is standard in pediatric neurocritical care, but the potential of non-invasive near-infrared spectroscopy (NIRS) and the bedside drivers of ICP remain under-explored. We evaluated the association between cerebral NIRS (rSO₂), pain and agitation, systemic hemodynamics, inflammatory markers, and invasive ICP in critically ill children. Methods: In this prospective pilot study, children undergoing external ventricular-drain-based ICP monitoring were followed with repeated measurements. Linear Mixed Models (LMM) with a random patient intercept were used to assess the NIRS–ICP association, adjusting for hemodynamic (mean arterial pressure (MAP), heart rate) and clinical variables. The Mann–Whitney U test compared parameters across a 20 mmHg ICP threshold, and the Kruskal–Wallis test compared ICP burden across etiology and outcome subgroups. Inflammatory markers (CRP, WBC, procalcitonin) were evaluated separately. Results: A total of 596 measurements from 21 patients were analyzed. Cerebral NIRS was the strongest independent predictor of ICP, showing a significant inverse association (estimate −0.609, p < 0.001). Agitation was independently associated with higher ICP (estimate 0.856, p < 0.001), and this effect was significantly modified by the patient’s behavioral state, being attenuated during sleep and calm wakefulness compared with agitated wakefulness (sleep × agitation interaction, p = 0.002). MAP retained an independent positive association with ICP (estimate 0.044, p = 0.019), whereas pain score did not. After adjustment for etiology and sex, these associations remained unchanged; etiology was itself an independent predictor of ICP (p = 0.001), whereas sex was not. Serum CRP independently predicted ICP (t = 2.382, p = 0.018), whereas WBC and procalcitonin did not. ICP burden differed by etiology, with intracranial masses and leukemia showing the highest loads (p < 0.001), but not across clinical-outcome groups (p = 0.192). Conclusions: Cerebral NIRS-derived rSO₂ was an independent inverse correlate of invasive ICP, and agitation was independently associated with ICP in a state-dependent manner, with its effect attenuated during sleep and calm wakefulness. CRP and etiology further shaped ICP burden. These findings support integrating non-invasive cerebral oximetry into multimodal pediatric neuromonitoring and highlight the value of managing distress to mitigate intracranial hypertension. Full article

Irrigation water management in arid systems requires spatially distributed estimates of crop evapotranspiration (ET) that fixed crop coefficients cannot provide. The actual ET of flood-irrigated rice (Oryza sativa L.) on the arid northern coast of Peru was mapped with the METRIC surface energy balance model (Mapping EvapoTranspiration at high Resolution with Internalized Calibration) on Google Earth Engine (GEE). Ten cloud-free Landsat 8/9 scenes (January–July 2022) were processed over 113 ha at Ferreñafe (Lambayeque) on the 30 m product grid, onto which the 100 m native thermal observation was resampled, with internal calibration based on automatic anchor-pixel selection and hourly ERA5-Land data. Daily field-mean ET ranged from 4.2 to 8.1 mm d⁻¹, peaking during flooding and establishment and declining towards harvest. Because the same reference $E{T}_o$ underlies the METRIC internal calibration and the FAO-56 estimate, this is a comparison between two modelling approaches rather than an independent validation. Against the FAO-56 reference ET, METRIC showed a positive bias of $+0.65$ mm d⁻¹ (percent bias (PBIAS) $=+13\%$; root mean square error (RMSE) $=1.23$ mm d⁻¹; $r^2=0.57$; $n=9$, after excluding one date with anomalous reanalysis forcing), concentrated during flooding and after harvest, whereas at full canopy cover the two estimates converged. Two global ET products that share neither the METRIC formulation nor the ERA5-Land forcing reproduce the same seasonal decline once the canopy closes ($r=0.63$ and $0.91$) but stay far below in magnitude, as expected from their 500 m pixel. ET did not differ between sowing methods and varied only slightly among cultivars (∼0.3 mm d⁻¹), against marked intra-field variability. The METRIC–GEE workflow offers a low-cost, high-resolution tool for monitoring water use in data-scarce arid rice systems. Full article

Background/Objectives: Multiple sclerosis (MS) is frequently accompanied by cognitive impairment, yet the neurobiological mechanisms underlying cognitive heterogeneity remain incompletely understood. The choroid plexus (CP), a blood–CSF barrier structure involved in cerebrospinal fluid production and neuroimmune signaling, has recently emerged as a potential MRI marker of inflammatory and neurodegenerative activity in MS. This scoping review mapped evidence on associations between CP features and cognitive functions in adults with MS. Methods: A broad search of PubMed, Scopus, Web of Science Core Collection, and the Cochrane Library identified 1163 records; after deduplication, screening, and full-text assessment, seven studies were included. Results: CP volume or normalized CP volume was assessed in all seven studies, and one study additionally examined the CP T1/T2 ratio. The Symbol Digit Modalities Test was used in six studies, while five applied multidomain cognitive assessment. Larger CP volume was associated in several studies with poorer baseline information-processing speed, visuospatial memory, or multidomain cognition, but longitudinal findings did not show consistent predictive value for CP volume. One study reported that a higher CP T1/T2 ratio predicted faster visuospatial-memory decline. Conclusions: CP-related measures may reflect broader neuroinflammatory and neurodegenerative processes relevant to cognition in MS, but the evidence remains limited and methodologically heterogeneous. Standardized acquisition and segmentation, harmonized cognitive assessment, and adequately powered longitudinal studies are needed to establish their independent and predictive value. Full article

Intelligent transportation and aerial remote sensing scenes suffer from complex scene variations, abundant miniature targets and unpredictable out-of-distribution obstacles, which brings tough mathematical challenges to open-world detection tasks. Conventional detection algorithms lack rigorous frequency-domain separation and spatial constraint mathematical formulations, resulting in severe tiny-object feature attenuation, inefficient multimodal feature matching and catastrophic forgetting during incremental category iteration. To solve these mathematical bottlenecks, this paper constructs the TPCA-Net model built upon frequency decomposition and spatial dependency mathematical modelling. The entire framework consists of four fixed core modules: High-Frequency-Aware Multi-Scale Feature Enhancement (HSE), Reparameterized Adaptive Text–Visual Alignment (RTA), Double Wildcard Spatial Dependency Fusion (WSF), and Incremental Forgetting-Free Dual-Path Detection (DPD). From the mathematical perspective, the HSE module adopts discrete cosine transform-based filtering equations to split high-frequency object details from low-frequency background signals and establishes cross-attention spatial constraint formulas to make up for missing contextual information of small targets. The RTA module introduces low-rank decomposition mathematical optimization and reparameterized tensor fusion rules to realize domain-adaptive text embedding calibration and zero-cost cross-modal mapping at the inference stage. The WSF module constructs dual-wildcard self-supervised mathematical loss to finish unsupervised unknown-object identification and builds decoupled semantic–spatial fusion equations to improve the positioning precision of novel targets. The DPD module designs two sets of independent optimization objective functions and category-freezing incremental mathematical constraints to avoid conflicting parameter updates and eliminate forgetting defects in new-class expansion. Validated on COCO, DOTA and AI-TOD datasets, TPCA-Net achieves 56.0% AP on COCO, 79.30% mAP on DOTA, and 40.5% overall AP with 28.7% small-object AP on AI-TOD while delivering an inference throughput of 101.2 FPS on the Tesla T4 edge GPU. The proposed method outperforms existing mainstream open-world detection algorithms in tiny-object and rare-category recognition while maintaining efficient inference speed. Full article

Animal re-identification (Re-ID) asks whether two images show the same individual, a recognition task that fits naturally into applications such as reuniting lost pets with their owners. Existing methods report strong scores, but typically under a single seed, one mask granularity, and no query-time corruption analysis, leaving open whether the gains survive deployment. We propose a hierarchical framework decoupling localisation (a YOLOv8 soft-crop) from identity embedding: a dual-stream network fusing a frozen CLIP ViT-B/16 (learned projection) with a fine-tuned ViT-Base carrying L2-norm part attention, trained under ArcFace. On a combined cat+dog open-set benchmark of 173 identities, it attains Rank-1 0.9742/mAP 0.8597 over three seeds, surpassing a ViT-only ablation by +2.39 Rank-1 and +2.07 mAP. Open-set verification shows all configurations converge near 68% true acceptance at the strictest false-acceptance rate. A background-bias evaluation brackets the embedding’s background reliance between a bounding-box lower bound and a SAM-silhouette upper bound; a manual audit retains the reliable soft crop. A nine-corruption audit identifies down-sampling and motion blur as dominant. On PetFace, the architecture retrieves across 14,716 unseen identities and remains viable in a few-shot regime. A Descriptor Vector Exchange (DVE) extension is Pareto-dominated, traced to the ViT’s coarse feature map and architectural redundancy. Full article

Multi-label classification of very-high-resolution remote sensing scenes is difficult not only because multiple land-cover categories coexist in one image, but also because their discriminative evidence is spatially uneven: boundaries, elongated structures, and fragmented regions are often weakened by appearance-dominated features; small categories can be suppressed by global scene responses; and large-area categories require broader spatial context. These spatial ambiguities are further complicated by label dependencies, where co-occurring categories may support each other while visually similar categories may compete under weak or incomplete local evidence. To address these coupled challenges, we propose SDRCNet, a lightweight structure-guided dual-relation consensus network for multi-label remote sensing scene classification. First, SDRCNet introduces structure-guided feature learning to strengthen boundary, directional, and regional structural cues while maintaining an efficient network design. Second, it learns label-aware query representations and aggregates multi-granularity evidence from global scene context, local detail responses, and regional patterns, enabling different categories to obtain evidence from suitable spatial scales. Third, an evidence-aware relation reasoning mechanism models both category correlations and category competitions, allowing the network to exploit supportive label context while suppressing conflicting predictions in complex scenes. Experiments on China-MAS-50k demonstrate that SDRCNet achieves 79.3% mAP and 83.7% OF₁. Under a comparable lightweight efficiency regime, it improves over RepViT-M1.1 by 1.9 and 0.7 percentage points in mAP and OF₁, respectively, while using fewer GFLOPs and parameters. Compared with the strongest official deep baseline, it improves both mAP and OF₁ by 5.3 points while using only 1.13 GFLOPs and 3.82M parameters. On MultiScene-Clean, SDRCNet further improves mAP from 64.8% to 70.5% and OF₁ from 71.3% to 75.9%, showing consistent effectiveness across different multi-label remote sensing benchmarks. Full article

The excessive cost and complexity of Industry 4.0 laboratory infrastructure limit the adoption of Digital Twin concepts in engineering education. This paper proposes a low-cost Digital Twin framework for sustainable manufacturing education integrating SAP NetWeaver, Node-RED, and AI-based decision support. The framework adopts a layered architecture that connects PLC-based simulation, IoT middleware, enterprise resource planning systems, and intelligent decision-making components. Node-RED enables real-time data exchange, while SAP NetWeaver provides enterprise-level integration through OData services. An AI module supports decision-making for production and inventory management. The framework has been validated through the implementation of a functional prototype and a series of end-to-end integration tests that evaluated communication reliability, system interoperability, API response performance, and AI-assisted decision-support capabilities. Competency-based mapping aligns the framework with Industry 4.0 engineering skills, supporting its use in academic environments. A sustainability assessment highlights reductions in infrastructure cost, energy consumption, and resource usage compared to traditional laboratory approaches. The results indicate that the framework has the potential to provide a scalable and accessible solution for teaching Digital Twin concepts, pending further classroom-based validation. Full article

Rice is one of the most important food crops in China, and frequent outbreaks of rice pests pose a serious threat to both the yield and quality of rice. Accurate monitoring of rice pests is therefore of great significance for field management and intelligent pest control. In this study, a refined rice pest detector, termed CLM-YOLO, is proposed based on YOLOv11n according to the characteristics of rice pest targets. Specifically, in the backbone network, the original C3k component in the C3k2 block is replaced with an improved RepViT-DBlock to form the C3k2R module, which strengthens early-stage local feature representation and improves the extraction of fine-grained pest-related cues. Additionally, the Multi-Dimensional Grouped Convolutional Block Attention Module (MDGCBAM) is embedded at the transition between the backbone and neck, allowing the network to emphasize pest-related regions while suppressing redundant responses from rice-field backgrounds. Finally, the Local Deformable Attention Adaptive Query Upsampling (LDAAQU) module is adopted in the neck to replace the original upsampling operation. Through deformable attention and query-guided adaptive aggregation, LDAAQU improves the spatial alignment of multi-scale features during feature fusion and enhances the recovery of fine-grained image details. Experimental results show that CLM-YOLO achieves favorable performance on key detection metrics. Specifically, the F1-score, mAP@0.5, and mAP@0.5:0.95 reached 84.37%, 87.8%, and 73.1%, respectively. The proposed method offers a feasible approach for accurate rice pest detection and intelligent pest monitoring in field environments. Full article