Search Results (7,005)

Fractures as critical discontinuous structural planes in rock masses, directly govern their stability and serve as the core controlling factor in rock mechanics engineering. Existing deep learning models for fracture extraction face persistent challenges, including imbalanced integration of deep and shallow features, limited suppression of background noise, inadequate multi-scale feature representation, and large parameter sizes—making it difficult to strike a balance between detection accuracy and deployment efficiency. Focusing on the Wanshanshan quarry in Yunnan, this study first constructs a high-precision digital model using close-range photogrammetry and 3D real-scene reconstruction. A lightweight yet high-accuracy intelligent detection method, termed MSEM-Deeplabv3+, is then proposed for rock mass fracture extraction. The model adopts lightweight MobileNetV2 as the backbone network, incorporating inverted residual modules and depthwise separable convolutions, resulting in a parameter size of only 6.02 MB and FLOPs of 30.170 G—substantially reducing computational overhead. Furthermore, the proposed MAGF (Multi-Scale Attention Gated Fusion) and SCSA (Spatial-Channel Synergistic Attention) modules are integrated to enhance the representation of fracture details and semantic consistency while effectively suppressing multi-source and multi-scale background interference. Experimental results demonstrate that the proposed model achieves an mPA of 89.69%, mIoU of 83.71%, F1-Score of 90.41%, and Kappa coefficient of 80.81%, outperforming the classic Deeplabv3+ model by 5.81%, 6.18%, 4.53%, and 9.2%, respectively. It also significantly surpasses benchmark models such as U-Net and HRNet. The method accurately captures fine and continuous fracture details, preserves the spatial distribution of long-range continuous fractures, and maintains robust performance on the CFD cross-scene dataset, showcasing strong adaptability and generalization capability. This approach effectively mitigates the risks associated with manual high-altitude inspections and provides a lightweight, high-precision, non-contact intelligent solution for fracture detection in high-steep rock slopes. Full article

Biodiversity monitoring in tropical ecosystems is constrained by limited infrastructure, insufficient localized datasets, and reliance on cloud-based tools. We introduce BEMAX, a lightweight convolutional neural network for offline classification of endangered tree species in the Bohol Biodiversity Complex, Philippines. A curated leaf-image dataset from five species and an unknown class was collected using a Raspberry Pi camera. The MobileNetV2-based model achieved a 93.89% validation accuracy and an 88.33% field accuracy. Deployed on a Raspberry Pi 4 with touchscreen and camera integration, BEMAX demonstrates embedded AI as a replicable framework for conservation in data-scarce environments. Full article

To address the challenges of steel surface defect detection—characterized by fine-grained textures, substantial scale variations, and complex background interference—conventional lightweight detectors often struggle to balance real-time navigation requirements with high-precision spatial localization on mobile inspection platforms. In this work, we propose KDM-YOLO, a lightweight visual localization and detection method built upon YOLOv10n, designed to provide an efficient perception engine for autonomous inspection robots. The proposed approach enhances the baseline through three key perspectives: feature extraction, context modeling, and multi-scale fusion. Specifically, KWConv is introduced to strengthen the representation of fine-grained texture and edge cues; C2f-DRB is employed to enlarge the effective receptive field and improve long-range dependency perception to reduce missed detections; and a multi-scale attention fusion (MSAF) module is inserted before the detection head to adaptively integrate spatial details with semantic context while suppressing redundant background responses. Ablation studies confirm that each module contributes to performance gains, and their combination yields the best overall results. Comparative experiments further demonstrate that KDM-YOLO significantly improves detection performance while retaining a compact model size and high inference speed. Compared with the YOLOv10n baseline, Precision, Recall and mAP@50 are increased to 91.0%, 93.9%, and 95.4%, respectively, with a parameter count of 3.29 M and an inference speed of 155.6 f/s. These results indicate that KDM-YOLO achieves an ideal balance between the accuracy and computational efficiency required for embedded navigation platforms, providing an effective solution for online autonomous inspection and real-time localization of steel surface defects. Full article

This study presents a comprehensive numerical investigation into the combined bending–shear behaviour of hollow-flanged cold-formed steel (HFCFS) beams filled with lightweight concrete (LWC). Although previous research has independently examined the pure bending and pure shear responses of these composite members, their structural performance under simultaneous bending and shear remains unexplored. In this work, advanced three-dimensional finite element (FE) models were developed in ABAQUS to simulate the nonlinear behaviour of LWC-filled HFCFS beams subjected to various shear-span ratios. The modelling approach was validated using published experimental data and extended through a systematic parametric study that considered three beam geometries, two steel yield strengths (350 MPa and 450 MPa), two lightweight-concrete strengths (30 MPa and 50 MPa), and aspect ratios ranging from 1.5 to 3.5. The results demonstrated a clear progression of governing failure modes, from web shear buckling at low aspect ratios to combined shear–flexure interaction at intermediate spans and flexural-dominated failure at larger spans. Normalised shear and bending demand–capacity ratios (V/Vu and M/Mu) were used to identify the dominant limit state, revealing a predictable transition from shear-controlled to flexure-controlled behaviour. The findings enhance the understanding of composite thin-walled steel–concrete systems under combined actions and highlight the need for dedicated design rules for CF-HFCFS beams operating within the bending–shear interaction domain. Full article

Wafer defect analysis is important for semiconductor manufacturing, but labeled data are limited, and class distributions are highly imbalanced. We present a semi-supervised framework with two lightweight hybrid CNN–Transformer models for wafer defect classification and segmentation. For classification, HybridCNN-ViT combines CNN-based local feature extraction with Transformer-based global context modeling, and adopts a three-stage progressive pseudo-labeling strategy to leverage unlabeled samples. The pseudo-label selection mechanism is systematically calibrated to improve pseudo-label reliability under limited labeled data. For segmentation, ConvoFormer-UNet integrates convolution-enhanced embeddings with Transformer blocks to balance boundary detail and global context. On the public WM-811K dataset, HybridCNN-ViT achieves 98.72% accuracy and 0.9985 macro-AUC under the semi-supervised setting for classification, while ConvoFormer-UNet reaches 99.19% IoU for segmentation with fewer parameters than several baselines. We also report efficiency on a single GPU to illustrate practical inference speed. Full article

We propose a coupling-aware equivalent source reconstruction framework for reproducing complex three-dimensional electromagnetic (EM) environments inside an anechoic chamber. A measured or simulated target field is represented by a finite set of physically realizable equivalent source antennas whose positions and complex excitations are identified by solving a nonlinear high-dimensional inverse problem. To ensure physical fidelity, the forward model explicitly accounts for mutual coupling through a full-wave Method-of-Moments (MoM) formulation, avoiding the inaccuracies of idealized uncoupled superposition. The inverse problem is efficiently solved using a hybrid evolutionary optimization scheme that combines an adaptive differential evolution strategy with stagnation-triggered CMA-ES refinement, augmented by a lightweight surrogate-based pre-screening to reduce expensive full-wave evaluations. The optimized source configuration is directly deployed in a microwave anechoic chamber, where the reconstructed field is measured on an observation plane and compared against the target field. The experimental results demonstrate close agreement in both amplitude and spatial distribution, while the proposed optimization pipeline substantially reduces the number of full-wave evaluations required for convergence. This work enables accurate repeatable chamber emulation of outdoor or in situ EM scenarios for robust system-level testing and evaluation. Full article

Emotion recognition plays a crucial role in human–computer interaction and mental health applications. Traditional Electroencephalogram (EEG)-based emotion recognition methods are limited in classification accuracy due to their neglect of the spatiotemporal characteristics of the signals and individual differences. This study proposes a novel EEG emotion recognition framework that integrates spatiotemporal features to enhance performance through the following innovations: (1) the use of a Recurrence Plot (RP) to transform one-dimensional EEG signals into two-dimensional images, enhancing the representation of nonlinear dynamic features; (2) the design of a Spatiotemporal Channel Attention Module (TCSA), which combines temporal convolution, channel, and spatial attention mechanisms to optimize the capture of complex patterns; and (3) the integration of the lightweight and efficient network Efficientnet to construct the TCSA-Efficientnet classification model. On the Database for Emotion Analysis using Physiological Signals (DEAP) dataset, the proposed method achieves accuracy rates of 99.11% and 99.33% for valence and arousal classification tasks, respectively. On the Database for Emotion Recognition Using EEG and Physiological Signals (DREAMER) dataset, the method achieves accuracy rates of 98.08% and 97.49%, outperforming other EEG-based emotion classification models on both datasets. This demonstrates its advantages in accuracy, robustness, and generalization. Full article

Accurate blueberry ripeness detection is crucial for intelligent harvesting but is challenged by complex orchard environments involving small, dense fruit clusters. This study proposes BBYOLOv12, an improved YOLOv12 model, to address missed detections and ripeness misjudgments. The method integrates a lightweight RepGhost backbone for efficient multi-scale feature extraction, a modified SimAM attention mechanism to enhance feature capture in dense regions, and an improved WIoU loss function to optimize small object localization. Evaluated on a self-built dataset, BBYOLOv12 achieved a mAP@0.5 of 98.97%, mAP@0.5:0.95 of 83.55%, precision of 97.55%, and recall of 97.27%, outperforming baseline and mainstream lightweight models. The model maintains high accuracy with only 2.36 million parameters and 5.59 GFLOPs, reducing complexity relative to the baseline. A practical Graphical User Interface was also developed for real-time detection and statistical analysis. This research provides an effective technical solution for multi-scale, dense perception tasks in agricultural applications. Full article

Anisogrid lattice structures are gaining increasing attention due to their high strength-to-weight ratios, which make them ideal for the production of lightweight mechanical components. This study presents a finite element model developed to evaluate stress distribution in an anisogrid sandwich panel with an octahedral core. The Taguchi method and analysis of variance (ANOVA) were employed to identify the geometric parameters that mostly influence the stress state and, consequently, the structural strength. The radius of the inclined ribs and the thickness of the skins were identified as the most critical factors, while the influence of the horizontal rib cross-sectional area was found to be minimal. The stiffness and strain energy of different cell geometries were also evaluated, and the results are consistent with the stress-based analysis. These findings offer valuable guidance for optimizing anisogrid geometry, improving load-bearing performance and advancing the design of high-efficiency structures. Full article

Background/Objectives: This study aims to develop a comprehensive Clinical Decision Support System (CDSS) that integrates multi-sequence sacroiliac joint (SIJ) MRIs with rheumatological, clinical, and laboratory findings into the decision-making process for the early diagnosis of axial spondyloarthritis (axSpA), incorporating segmentation-supported explainability. Methods: Multi-sequence SIJ MRI data (T1-WI, T2-WI, STIR, and PD-WI) were analysed from 367 participants (n = 193 axSpA; n = 174 non-axSpA controls). Sequence-based classification was performed using VGG16, ResNet50, DenseNet121, and InceptionV3 models; additionally, a lightweight and parameter-efficient SacroNet architecture was developed. Slice-level probability scores were converted to patient-level scores using the Dynamic Top-K Averaging method. Image-based scores were combined with a logistic regression-based clinical risk score using weighted linear integration (0.60 image/0.40 clinical) and a conservative threshold (τ = 0.70). Grad-CAM was applied for visual interpretability. Furthermore, to support the diagnostic outcomes with precise spatial data, active inflammation in STIR and T2-WI sequences was segmented. For this purpose, the MDC-UNet model was employed and compared with baseline U-Net derivatives. Results: Sequence-specific analysis showed VGG16 performing best on T1-WI (AUC = 0.920; Accuracy = 0.878) and DenseNet121 on STIR (AUC = 0.793; Accuracy = 0.771). The SacroNet architecture provided competitive classification performance at the patient level despite its low number of parameters (~110 K). Furthermore, MDC-UNet successfully segmented active inflammation, yielding Dice scores of 0.752 (HD95: 19.25) for STIR and 0.682 (HD95: 26.21) for T2-WI. Conclusions: The findings demonstrate that patient-level decision integration based on multi-sequence MRI, when used in conjunction with clinical risk scoring and segmentation-assisted interpretability, can provide a feasible and interpretable DSS framework for the early diagnosis of axSpA. Full article

Topology optimisation is a powerful methodology for identifying efficient material distributions within prescribed design domains. However, conventional approaches rely heavily on gradient-based optimisation and repeated numerical simulations, which impose significant computational cost and limit their use in early-stage design exploration. This work introduces a generative design framework, referred to as Large Language Model-Driven Symbolic Topology Optimisation (LLM-DSTO), in which large language models act as conceptual design generators. Engineering problems are formulated through structured textual descriptions defining the design domain, boundary conditions, loading scenarios, and material constraints. The language model interprets these inputs and produces symbolic representations of candidate structural topologies. The generated layouts are evaluated using physics-informed objective functions and refined iteratively through lightweight computational procedures. The resulting designs exhibit coherent load paths, strong structural connectivity, and material distributions that are consistent with practical manufacturing requirements, including additive manufacturing constraints. The proposed framework is validated across structural, thermal, thermofluid, and compliant mechanism design problems. Quantitative results show that the generated structures achieve approximately 87.5% of the stiffness obtained using the classical SIMP method for the cantilever benchmark, while reaching about 94.3% of the thermal performance in heat sink optimisation. These results are obtained without repeated finite element simulations, demonstrating a significant reduction in computational cost. In addition, the framework is extended to three-dimensional topology generation, producing volumetric structures under a 50% material volume constraint with coherent internal load paths. Full article

The freeze–thaw cycles lead to cumulative damage and gradual strength deterioration in concrete, which cannot be accurately represented by traditional empirical models. To address this issue, shale-based lightweight structural concrete (LSC) specimens with four strength grades (LSC20-LSC50) were subjected to basic mechanical performance and freeze–thaw cycle experiments. The study investigated the patterns of mass loss, relative dynamic elastic modulus, and loss of compressive strength in LSC subjected to varying numbers of freeze–thaw cycles. Furthermore, the correlation between the dynamic modulus of elasticity and compressive strength was examined. A Gamma process-based stochastic degradation model for the compressive strength of LSC was then developed. The results show that the compressive strength degradation of LSC under freeze–thaw cycles follows a monotonically increasing trend that gradually stabilizes, with low-strength LSC deteriorating faster than high-strength LSC. After 200 cycles, the compressive strength degradation of LSC30 and LSC50 was only 32.66% and 29.79% of that of their corresponding ordinary concretes (C30 and C50). The proposed Gamma process model showed high fitting accuracy for all strength grades of LSC (R² > 0.96, RMSE < 0.25 MPa, MAPE < 11%). The research results provide a scientific basis for the structural design of concrete in cold regions. Full article

Cross-view geo-localization tries to find the matching place in large satellite or aerial pictures from photos taken at ground level, which is useful for applications like self-driving cars, flying drones, and adding virtual objects to real city scenes. However, the traditional deep learning hybrid CNN-Transformer architecture and complex geometric submodules result in a large computational overhead, making it difficult to apply in real-time on resource-constrained devices. To make it light, fast, and accurate, this paper suggests an effective way to make a state-space model for cross-view geo-localization tasks. The model replaces the traditional self-attention structure with a state-space vision backbone, lowering the sequence modeling complexity from quadratic to linear and greatly accelerating the inference process; it devises a channel-group aggregation strategy without any learnable parameters, producing a comprehensive yet lightweight representation, and introduces a dynamic difficulty-aware loss function that assigns varying weights to all negative samples within a batch according to their similarities, greatly improving the efficiency of hard-negative sample mining and the quality of convergence. The results on the authoritative public datasets CVUSA and CVACT indicate that our method has high accuracy and low computational complexity, providing a feasible approach for the lightweight design of more powerful cross-view geolocation models in the future. Full article

In low-altitude complex environments, accurately reconstructing the three-dimensional (3D) flight trajectories of small unmanned aerial vehicles (UAV) without onboard positioning modules remains challenging. To address this issue, this paper proposes a multi-view ground camera-based UAV 3D trajectory detection method founded on an optical imaging geometric model. Multiple ground cameras are used to synchronously observe UAV flight, enabling stable 3D trajectory reconstruction without relying on onboard Global Navigation Satellite System (GNSS). At the two-dimensional (2D) observation level, a lightweight object detection model is employed for rapid UAV detection. Foreground segmentation is further introduced to extract accurate UAV contours, and geometric centroids are computed to obtain precise image plane coordinates. At the 3D reconstruction stage, camera extrinsic parameters are estimated using a back intersection method with ground control points, and the UAV spatial position in the world coordinate system is recovered via multi-view forward intersection. Field experiments demonstrate that the proposed method achieves stable 3D trajectory reconstruction in real urban environments, with a median error of 4.93 m and a mean error of 5.83 m. The mean errors along the X, Y, and Z axes are 2.28 m, 4.58 m, and 1.09 m, respectively, confirming its effectiveness for low-cost UAV trajectory monitoring. Full article

Generating high-precision normalized digital surface models (nDSMs) from a single remote sensing image remains a challenging and ill-posed problem due to the absence of reliable geometric constraints. In this work, we show that monocular depth provides structurally stable cues of local geometry but lacks the global scale and vertical reference required for absolute height recovery. This intrinsic mismatch limits direct depth-to-height regression, particularly when transferring across heterogeneous terrains, land-cover compositions, and imaging conditions. Building on this idea, we propose the Relative Depth–Absolute Height Prediction Network (RDAH-Net), a framework that exploits relative depth as a geometry-aware prior while learning terrain-dependent height mappings from image appearance to absolute height. As the backbone, we employ a lightweight MobileNetV2 enhanced with a Convolutional Block Attention Module (CBAM), and further incorporate a cross-modal bidirectional attention fusion scheme with positional encoding to achieve a deep and effective fusion of image appearance and depth prior cues. Finally, a PixelShuffle-based upsampling strategy is used to sharpen prediction details and mitigate typical upsampling artifacts. Extensive experiments across diverse regions demonstrate that RDAH-Net achieves robust and generalizable height estimation, providing a practical alternative for large-scale mapping and rapid update scenarios. Full article