Search Results (10,197)

There have been significant advancements in object detection using extensive labelled datasets. However, existing learning-based approaches remain constrained in industrial environments, primarily due to the limited diversity in training datasets; the lack of generalisation of close-set detectors to unseen asset categories; and the inherent spatial and geometric complexity of mechanical, electrical, and plumbing (MEP) assets. To address this challenge, we propose a new approach that leverages pre-trained vision language models and close-set object detectors to detect unseen MEP assets using unlabelled data. Experimental results reveal the superior performance of Grounding DINO using Swin B transformer in open-vocabulary MEP asset detection, achieving the mean intersection over union (mIoU) of 0.6586 for valve detection and 0.4883 for pump detection. In addition, the combination of Grounding DINO (Swin B) and YOLOv8 outperforms other configurations in MEP asset detection, attaining the highest performance for both valve detection, with mean average precision at IoU = 0.5 (mAP50) of 0.928 and mean average precision over IoU threshold from 0.5 to 0.95 (mAP50:95) of 0.889, and pump detection, with corresponding values of 0.778 and 0.662, respectively. The quantitative and qualitative results of our approach were evaluated against fine-tuned Grounding DINO and fully supervised close-set object detectors. Full article

Automated pavement distress classification using high-resolution Unmanned Aerial Vehicle (UAV) imagery is pivotal for intelligent transportation systems. However, long-term UAV monitoring faces a continuous stream of evolving distress types and changing remote sensing background textures, necessitating Class-Incremental Learning (CIL) capabilities. Existing methods struggle to balance stability and plasticity, especially under the severe storage limitations typical of local edge stations in air–ground collaborative systems. This data scarcity leads to catastrophic forgetting and confusion among fine-grained distress categories. To address these challenges, we propose a data-efficient approach named Analytic Geometric Alignment (AGA). Our framework mainly consists of three key components. First, to overcome the optimization gap between the feature extractor and the fixed geometric target, we introduce a Subspace-Aware Analytic Initialization (SAI) that computes a closed-form projection to instantly align the feature subspace with the ETF manifold before each task training. Second, on this aligned basis, a Decoupled Geometric Adapter (DGA) is incorporated to facilitate continuous non-linear adaptation to complex aerial textures. Finally, for stable incremental training, we design a Memory-Prioritized Regression (MPR) loss to enforce tighter geometric constraints on replay samples, significantly enhancing model stability. Extensive experiments on the UAV-PDD2023 dataset demonstrate that AGA significantly outperforms state-of-the-art methods, showcasing excellent robustness and data efficiency. Full article

Onboard GNSS-based relative baseline determination has emerged as a primary solution for formation-flying satellites dedicated to mapping and remote sensing missions. For ambiguity resolution (AR), the double-difference (DD) method is widely adopted in relative baseline determination. However, this method entails relatively complex satellite pairing, which not only increases computational load and complicates the processing workflow but also imposes higher requirements on onboard embedded computing and storage resources, thereby introducing potential risks to engineering implementation. To address these issues, this paper proposes incremental refinements to the single-difference (SD) model by introducing the combined GPS/BDS uncombined SD method for closely spaced formation satellites. By leveraging the enhanced satellite visibility of the combined GPS/BDS constellation and adopting a purely geometric approach, high-precision real-time relative baseline determination results are achieved. Validation using onboard observation data from the Lutan-1 satellite mission of China demonstrates that centimeter-level relative baseline determination accuracy can be attained. Full article

Monitoring surface subsidence in mining areas is essential for geological disaster early warning and safe production. Existing geometric difference methods heavily rely on the local consistency of multi-temporal point clouds. When horizontal displacement and vertical subsidence are coupled, horizontal movements often cause local misalignments, leading to spatial deviations and discrete anomalies in vertical estimations. To address this issue, this paper proposes DL-C2C, a deep learning model for subsidence extraction from bi-temporal ground point clouds. Within a unified framework, the model introduces horizontal displacement as an auxiliary constraint into the vertical solving process, effectively improving the stability of vertical subsidence estimation through continuous cross-temporal alignment and correlation updating. For feature extraction, DL-C2C employs a PointConv multi-scale pyramid combined with a proposed scale-adaptive Transformer to enhance cross-scale information interaction under sparse and non-uniform sampling conditions. Furthermore, the network constructs dynamic local associations through iterative alignment within a recursive framework, and introduces diffusion-based residual correction at the fine-scale stage to compensate for detail errors at subsidence basin boundaries and in data-missing regions. Experiments on simulated and real-world datasets—covering aeolian sand and mountainous gully landforms—demonstrate that the method achieves mining 3D error (M3DE) of 0.16 cm and 0.22 cm in simulated scenarios. In real-world mining area validations, compared to existing methods, DL-C2C significantly reduces discrete anomalous points, yields an error distribution closer to zero, and exhibits superior performance in boundary transition continuity and non-subsidence area stability. In conclusion, this model provides reliable technical support for large-scale, high-precision intelligent monitoring of geological disasters in mining areas. Full article

Accurate 3D documentation of large and complex structures is essential for long-term stability assessment, structural monitoring, and conservation planning, particularly for heritage sites exposed to environmental and anthropogenic threats. This study develops an integrated workflow combining Terrestrial Laser Scanning (TLS), Global Navigation Satellite System (GNSS), and Total Station geodetic control for large-scale, high-precision documentation. The approach was implemented at the Saqqara archaeological zone, a UNESCO World Heritage Site facing significant deterioration risks, to document four major pyramids: Djoser, Unas, Teti, and Userkaf. More than 2.1 billion georeferenced points were acquired from 16 scan positions with sub-centimeter registration errors and overall geometric accuracy better than ±1 cm. From these datasets, detailed mesh models, orthoimages, Digital Elevation Models (DEMs), contour maps, and 2D plans were derived. These enabled quantitative analyses of height loss and volumetric change, indicating severe structural degradation in Unas (~53%), Teti (~66%), and Userkaf (~63%), as well as localized deformations such as 4.2 cm displacement at Teti’s south flank. The degradation results from environmental factors and anthropogenic influences. Beyond this case study, the workflow proves that integrated TLS documentation can be applied to large and complex structures, supporting deformation monitoring, stability assessment, and digital twin development. Full article

During the high-temperature deposition of tungsten thin films on alumina ceramic substrates, the inherent mismatch in thermal expansion coefficients frequently triggers interfacial delamination, where uncontrollable factors in stochastic surface topographies can exacerbate localized stress concentrations. To resolve these interfacial failures, the enhancement of interfacial adhesion through a deterministic surface microgroove design is identified as the general objective of the present research. Within this framework, the establishment of a robust quantitative mapping between the transverse scratching offset distances and the resultant periodic microgeometry is first pursued as a specific experimental objective. This methodological approach effectively transforms the stochastic nature of the substrate into deterministic geometric configurations. Second, a specific numerical objective is fulfilled by evaluating the interfacial stress redistribution and damage evolution utilizing refined thermomechanical coupled simulations based on the cohesive zone model. The integrated findings demonstrate that optimizing the microgroove spacing effectively governs the morphological transition and broadens stress diffusion pathways to mitigate thermal mismatch effects. Specifically, the structural optimization at a spacing of 28.8 μm facilitates an approximately 31.8% reduction in the maximum interfacial stress and a 10% decrease in the average film stress compared to the 13.6 μm spacing. Finally, this research clarifies the underlying mechanisms of stress buffering and provides a rigorous engineering methodology for the structural design of reliable high-performance ceramic–metal interfaces in extreme environments. Full article

With the increasing functional and geometric complexity of modern steel buildings, irregular beam-to-column joints are becoming increasingly common in engineering practice, while their seismic performance and force transfer mechanisms remain insufficiently understood. Based on previous full-scale cyclic loading tests on unequal-depth steel beam (UDSB) and staggered steel beam (SSB) joints incorporating inclined internal diaphragms, this study presents numerical simulations and parametric analyses of irregular steel beam to concrete-filled steel tube (CFST) column joints. Three-dimensional nonlinear finite element models were developed using ABAQUS and validated against experimental results. The strengthening effects of internal diaphragms and concrete infill were then comparatively investigated. The results indicate that internal diaphragms increase the initial stiffness and load-carrying capacity of the joints to approximately 2.0–2.3 times and 1.16–1.8 times, respectively, compared with joints without diaphragms, whereas concrete infill provides smaller enhancements of about 1.3 times in stiffness and 1.2–1.3 times in strength. In addition, the hysteretic response of joints without diaphragms shows good agreement with the post-fracture behavior observed in the experiments, validating the diaphragm fracture mechanism. A parametric study further demonstrates that, under cyclic loading, the beam depth ratio, staggered floor ratio, column wall thickness, column width, diaphragm thickness, and diaphragm opening diameter have significant influences on joint strength and stress distribution, while the effect of axial load ratio is relatively minor. Finally, a strength prediction method applicable to inclined-diaphragm UDSB and SSB joints is proposed, and corresponding fitted expressions are derived based on the parametric results. The findings provide useful guidance for the seismic design of irregular steel beam–CFST column joints incorporating internal diaphragms. Full article

(1) Background: Curve data compression plays a critical role in efficient storage, transmission, and multi-scale visualization of vector spatial data, especially for complex geographic boundaries. Achieving high compression efficiency while preserving geometric fidelity remains a challenging task. (2) Methods: This study proposes a vector curve compression framework based on a convolutional autoencoder. Curve data are segmented and resampled to unify network input, after which coordinate-difference sequences are encoded into low-dimensional latent vectors through convolutional layers and reconstructed via a symmetric decoder. (3) Results: Experiments conducted on a global island boundary dataset demonstrate that the proposed method achieves effective data reduction with stable reconstruction accuracy. Specifically, compared with the classical Douglas–Peucker (DP) algorithm, Fourier series (FS) methods, and fully connected autoencoders (FCAs), the 1D CAE exhibits superior and more robust reconstruction performance, especially under high compression ratios. It achieves the lowest positional deviation (PD = 42.41) and the highest spatial fidelity (IoU = 0.9991, with a relative area error of only 0.0067%), while maintaining high computational efficiency (57.32 s). Sensitivity analyses reveal that a convolution kernel size of 1 × 7 and a segment length of 25 km yield the optimal trade-off between representational capacity and model stability. (4) Conclusions: The proposed method enables efficient vector curve compression and reliable coastline reconstruction, and is particularly suitable for small- and medium-scale cartographic applications up to a map scale of 1:250 K. Full article

Accurate prediction of cutting force is essential for process optimization and intelligent control in CNC turning, yet cross-material performance comparisons of machine learning models remain limited. This study develops and applies a structured diagnostic benchmarking framework to evaluate ten supervised regression models for cutting force prediction across five engineering alloys: Aluminum Alloy 6061, Brass C26000, Bronze C51000, Stainless Steel 304 (annealed), and Carbon Steel 1020 (annealed). The input space included material category together with machining descriptors (diameter, feed rate, and axial distance from the chuck). Model performance was evaluated using the coefficient of determination (R²), root mean square error (RMSE), cross-validated stability metrics, and pairwise dominance probability matrices derived from R² and CV(RMSE). Gradient Boosting achieved the highest overall accuracy and robustness, with a mean R² = 0.962 and RMSE = 18.03 N, followed by a feedforward neural network (R² = 0.953, RMSE = 19.96 N), while Support Vector Regression showed substantially lower performance (R² < 0.65; RMSE > 54 N). Residual diagnostics indicated that ensemble and neural models produced compact, near-homoscedastic error distributions, whereas linear and single-tree models exhibited systematic bias and heteroscedasticity. Principal Component Analysis revealed that the first two components captured 78.7% of the total variance, separating geometric and spatial effects from feed-driven variability. The proposed evaluation framework provides a unified methodology for accuracy, and multivariate interpretation in machining force prediction. These results offer practical guidance for selecting robust learning models in intelligent CNC systems and data-driven manufacturing environments. Full article

Machining stability in profile grinding directly affects surface quality and form accuracy, while the variation in local contact conditions induced by complex contour geometries makes its stability behavior more complicated than that of conventional grinding. This study investigates chatter stability under the coupled effects of contour geometric features and process parameters. A dynamic grinding force model is developed based on a tool nose micro-element method, explicitly considering the coupled effects of contour geometric parameters, wheel–workpiece contact, and regenerative effects. A chatter stability model is then established, and an iterative method is proposed to predict stability limits under different contour features. The results indicate that wheel speed and grinding depth dominate system stability. Under the same curvature radius, convex contours exhibit the highest stability, followed by straight and concave contours. As the curvature radius increases, the stability boundaries gradually converge toward that of the straight contour. Increasing the contour normal angle (CNA) significantly enhances stability and promotes the transition of the dominant unstable mode from single-direction to multi-directional coupling. Grinding experiments on a composite curved workpiece validate the model, showing strong agreement between predicted stability regions and measured chatter marks and spectra. The proposed model provides a basis for parameter selection and chatter suppression in complex profile grinding. Full article

This paper presents a methodology enabling the use of a Pitot probe for static pressure measurement in supersonic flow under severely confined spatial conditions where standard design guidelines cannot be satisfied. In particular, the recommended placement of a static pressure tapping at a distance of 10–20 tube diameters is not feasible; the proposed approach allows for the tapping to be located immediately downstream of the static tube cone. The methodology combines theoretical analysis, experimental measurements, and Computational Fluid Dynamics (CFD) simulations. Experiments were performed using appropriately selected pressure sensors, while detailed simulations in Ansys Fluent (Ansys 2024 R2) included both a high-fidelity probe model and free-stream flow analysis. By comparing experimental and numerical results, a correction coefficient was established based on the free-stream dynamic pressure obtained from CFD. This enables the accurate estimation of static pressure even in non-ideal probe configurations. The measurement error did not exceed 20%, while in most cases, very good agreement between experimental and CFD results was achieved. The main contribution of this paper is the validated methodology, which extends the applicability of Pitot probes to geometrically constrained environments where conventional static pressure measurement techniques cannot be implemented. Full article

Skeleton-based action recognition is a critical technology for intelligent sports analysis. Although the human skeletal structure exhibits inherent bilateral symmetry, sensor noise on resource-constrained edge devices frequently induces geometric distortion and topological asymmetry. Consequently, achieving a balance between high accuracy and real-time performance remains a significant challenge. To this end, we propose EMS-GCN, an Efficient Multi-scale Shift Graph Convolutional Network that integrates geometric priors. Specifically, we design a Gaussian kernel-driven topology refinement module to mitigate structural noise inherent in sensor data. By leveraging geometric symmetry and Gaussian distances among nodes, this module dynamically constrains graph topology learning, thereby effectively rectifying the structural asymmetry and ambiguity induced by noise. Furthermore, we construct a Multi-scale Shift Linear Attention (MSLA) module to replace computationally intensive temporal convolutions. Leveraging temporal shift invariance, this module captures multi-scale contexts via parameter-free shift operations. Furthermore, we introduce a linear temporal attention mechanism to model global temporal dependencies with linear complexity, effectively resolving the information asymmetry inherent in long-range interactions. Finally, EMS-GCN incorporates a dual-branch attention structure to adaptively calibrate feature responses. Extensive experiments demonstrate that our model maintains high recognition accuracy with only 0.56M parameters, representing a reduction of over 60% compared to mainstream baselines. These results validate the efficacy of leveraging geometric and temporal symmetries to enhance real-time sports analysis. Full article

This study investigates the residual stresses developed during the curing process of polymer fiber-reinforced composites and their influence on fracture behavior, particularly the initiation and propagation of interlaminar cracks. The main objective is to quantify how different curing histories, including incomplete cure, alter the spatial distribution of residual stresses and, in turn, affect the mode-I fracture response of carbon-fiber/epoxy laminates. A transient thermal–structural finite element framework incorporating an autocatalytic cure kinetics model was used to simulate the curing process and predict residual stress development in a unidirectional carbon-fiber/epoxy laminate with an edge crack, considering thermal, chemical, and geometric effects. The cure model was calibrated using isothermal differential scanning calorimetry data to determine the degree of cure under different thermal conditions. The key novelty of this work is the integration of a validated cure-kinetics-based curing simulation with fracture analysis, enabling direct correlation of thermal history and degree of cure with spatially varying residual stresses at the crack front and their effect on fracture toughness. Numerical load–displacement predictions were compared with double cantilever beam experimental results and showed good agreement for the curing profiles examined. The results demonstrate that residual stresses generated by different cure cycles, including hold conditions and incomplete curing, significantly influence fracture toughness. In particular, the incomplete-cure profile produced an approximately 40% reduction in toughness compared with profiles that achieved complete cure, highlighting the importance of cure history in determining final structural performance. Full article

This study proposes an innovative structural wall system and evaluates its compressive performance. The wall consists of GLPB manufactured using laminated bonding (along the grain direction) and assembled using a staggered interlocking masonry method. Two key geometric parameters controlling the mechanical response of the GLPB wall—the slenderness ratio (β) and the eccentricity (e)—were selected as the primary design variables. Using a combined experimental and numerical approach, the study systematically investigated the compressive mechanical behavior and performance evolution of the wall, including compressive strength and deformation behavior. Through axial and eccentric compression tests, six sets of specimens with varying geometric parameters β and e were analyzed, yielding relevant data and characteristics regarding failure modes, ultimate load-carrying capacity, load–displacement response, crack resistance, and wall deformation. To further characterize the compressive mechanical performance of GLPB walls, a refined nonlinear finite element model was developed in ABAQUS (version 2020). This model incorporates the anisotropic constitutive behavior of wood, the Hill yield criterion, and the mechanical interactions at the interlocking and bonding interfaces. The study indicates that the average compressive strength of GLPB walls is 2.63 MPa, with a crack-to-failure load ratio ranging from 0.68 to 0.83. GLPB walls demonstrate comparable load-bearing capacity. The total axial vertical strain ranges from 0.033 to 0.041, indicating that the walls possess good deformation capacity. Based on Chinese masonry design standards and experimental evidence, a preliminary predictive formula for the load-bearing capacity of this wall was derived. A comparison of the aforementioned experimental measurements with simulation results showed errors of less than 10%, verifying the model’s validity and accuracy. Numerical simulation can, to a certain extent, compensate for the limitations of experimental methods in capturing internal mechanical states. Full article

Optical-fibre daylighting systems (OFDS) harvest solar energy as a renewable lighting resource by delivering sunlight deep into green buildings. This emerging technology for sustainable infrastructure reduces electric-lighting demand; however, reported performance is difficult to compare across heterogeneous designs, metrics, and validation practices. Therefore, a PRISMA 2020–reported systematic literature review (SLR) of OFDS studies from three databases (Google Scholar, Scopus, and Web of Science; 2000–2025) was conducted, synthesising primary research that quantifies system- or component-level performance, with a focus on (i) plastic optical fibre (POF) transmission characteristics; and (ii) POF-based illuminance model validation. After de-duplication and screening, 106 primary studies were included, and two meta-analyses were performed where data were harmonised from 29 studies in total. Across reported POF configurations, attenuation ranged from 150 to 800 dB/km with a pooled mean of 332.8 dB/km, corresponding to a mean 1 m transmission of 92.7% and median design length scales of ∼3.7 m for 80% transmission and ∼11.6 m to half-power. Across illuminance validation datasets, models showed high linear agreement with experimental measurements (coefficient of determination ($R^2$) = 0.99; slope = 0.99) but typically underpredicted illuminance (geometric mean ratio = 1.16; mean absolute error (MAE) = 27.3 lux; mean absolute percentage error (MAPE) = 17.6%). These findings underscore the need for a standardised evaluation framework, including consistent metric definitions, robust uncertainty reporting, and reusable validation datasets to enable variance-weighted synthesis, while also identifying short-run POF routing as a key lever for improving system efficiency. In addition to providing the OFDS research agenda, this study serves as a roadmap for the industrial development of daylighting systems for green buildings based on harvesting solar energy, with its novelty lying in the PRISMA-guided evidence synthesis and quantitative meta-analytic consolidation of POF transmission and illuminance-validation performance. Full article