Search Results (625)

Depth super-resolution reconstructs high-resolution (HR) depth maps from low-resolution (LR) inputs with the aid of HR RGB guidance, but RGB edges often do not coincide with true depth discontinuities, causing texture copying and degraded geometric consistency. To address this problem, we propose Frequency–Geometry-Guided Network (FGGNet), a spatial–frequency fusion framework for RGB-guided depth map super-resolution. FGGNet introduces Multi-branch RGB-guided Convolution (MRGConv) to enhance RGB structural representations, a Geometry Prior-guided Fusion Module (GPFM) to filter geometrically inconsistent RGB responses using depth-derived priors, and radial complex spectral loss (RCSL) to emphasize boundary-related high-frequency components in the complex spectral domain. Experiments on NYU v2, Middlebury, Lu, and RGB-D-D show that FGGNet achieves competitive or superior reconstruction accuracy under synthetic and real-world degradation settings. Under the ×16 setting, FGGNet reduces RMSE by 13.7%, 22.8%, 18.5%, and 11.4% on the four datasets, respectively, compared with the average RMSE of five representative state-of-the-art methods. These results validate the effectiveness of combining geometric prior filtering with frequency-domain supervision for reliable depth reconstruction. Full article

LiDAR point clouds are widely used in remote sensing perception scenarios, such as autonomous driving. However, LiDAR-based perception models remain vulnerable to adversarial perturbations, which may compromise the reliability of safety-critical 3D perception systems. Among different attack paradigms, transfer-based attacks are particularly practical because they generate adversarial examples on accessible surrogate models and apply the generated examples directly to unknown target models. Nevertheless, existing transferable attacks on point clouds often perturb regions that are discriminative for the surrogate model but insufficiently stable across different architectures, leading to limited transferability and noticeable geometric distortion. To address this problem, we propose SAGE, a Saliency And Geometry Enhanced transferable attack framework for LiDAR point cloud perception in remote sensing. Specifically, SAGE unifies point-coordinate priors with source-model gradient signals to generate a saliency map, which serves as a transferable indicator of vulnerable local structures. SAGE further leverages this map through saliency-guided perturbation allocation and explicit geometric constraints to enhance transferability while preserving point-cloud geometry. To demonstrate the effectiveness of SAGE, we evaluate SAGE on point-cloud classification benchmarks and further validate it on LiDAR-based 3D object detection using KITTI and nuScenes. Experimental results show that SAGE consistently outperforms existing transferable attack methods in attack success rate while preserving favorable geometric quality of adversarial point clouds. These findings demonstrate that SAGE offers an effective and practical framework for assessing the transfer robustness of LiDAR-based remote sensing perception systems. Full article

Precise segmentation of high-magnification titanium alloy micrographs under few-shot scenarios remains a non-trivial task, primarily owing to the intricate morphology, heterogeneous discrete distribution, and weak phase boundaries of the primary α phase. To address these issues, this paper presents MaterialAlphaSAM, a lightweight domain-adaptive segmentation framework built upon the Segment Anything Model (SAM). Leveraging SAM’s powerful global context modeling capability, the proposed method incorporates two key modules: a Geometry-Constrained Prompt Prior (GCPP) module and a Domain-Adaptation Adapter (DAA) module. The GCPP module explicitly embeds geometric and morphological priors to generate semantically guided prompts, effectively alleviating prompt redundancy and noise sensitivity. The DAA module performs cross-domain alignment of the encoder features, reducing the domain discrepancy between natural images and metallic microstructures. Extensive experiments demonstrate that both modules consistently boost segmentation performance. On the titanium alloy dataset, MaterialAlphaSAM achieves 89.53% IoU and a 94.40% F1-score, outperforming FCN, UNet, DeepLabV3, PSPNet and the vanilla SAM. It exhibits superior robustness to weak boundaries, fine-scale α phases, and complex background interference. Full article

Accurate perception and 3D reconstruction of fruit tree branch structures are fundamental to smart orchard development, with broad applications in intelligent harvesting, crop phenotyping, and precision management. However, the slender and highly branched morphology, multi-scale distribution, weak surface texture, and severe occlusion inherent to fruit tree branches pose substantial challenges to high-fidelity modeling. This paper systematically reviews advances in branch feature extraction and 3D reconstruction for fruit tree canopies. A structured literature search was conducted using the Web of Science, Scopus, and Google Scholar databases, with search terms including “fruit tree branch”, “point cloud reconstruction”, “3D canopy modeling”, “branch feature extraction”, and “agricultural robotics”. Studies published between 2000 and 2025 were considered, with inclusion criteria requiring relevance to branch structure perception, reconstruction accuracy, or orchard application; non-peer-reviewed sources and studies lacking quantitative evaluation were excluded. We trace the evolution of feature extraction from classical 2D image processing and geometric fitting, through point cloud segmentation and skeleton extraction, to modern deep learning approaches and multimodal perception techniques. For 3D reconstruction, we compare active and passive sensing strategies alongside both explicit and implicit scene representation methods, discussing their respective strengths and applicable scenarios. A five-dimensional evaluation framework is also proposed, encompassing geometric accuracy, structural consistency, feature stability, computational efficiency, and generalization capability. Finally, we identify key bottlenecks in fine-grained structure recovery, occlusion handling, and cross-scene generalization, and highlight future directions in structural prior integration, multimodal collaborative modeling, and lightweight neural representations—offering a structured reference for advancing 3D perception research in smart orchards. Full article

Citrus peel is examined here as a naturally evolved protective surface, with the goal of developing a transferable quantitative framework for extracting engineering-relevant descriptors from biological protective surfaces and using them as design templates for biomimetic counterparts. A single-specimen-per-species design is adopted to map intra-fruit geometric variation across regions and magnifications; absolute descriptor values are therefore reported as ordinal indicators of inter-species ranking rather than as population means. Five citrus species (lemon, orange, mandarin, grapefruit, and bitter orange) were characterised by mechanical testing (cutting, puncture, and compression; five replicates per fruit), gravimetric peel density and thickness, and scanning electron microscopy (SEM) at 100×–10,000×. The 135-image SEM dataset was processed with an automatic-calibration pipeline performing per-image scale-bar detection, multilevel-Otsu segmentation of albedo air space, cell-bounded surface segment (CBSS) and oil-gland segmentation on flavedo, and grey-level co-occurrence matrix (GLCM) texture analysis with a directional anisotropy index $A_F$. Calibration was consistent across all images (FoV × magnification $=\mathrm{403,273}\pm 410$ $\mathsf{\mu }$m$·\times$, $\pm 0.10\%$). Principal component analysis separated flavedo and albedo at every magnification (PC1 + PC2 = 84–92%). Within this dataset, grapefruit showed the densest CBSS cover (1072 ${\mathrm{mm}}^{-2}$) together with the highest oil-gland density (2.77 ${\mathrm{mm}}^{-2}$); bitter orange showed the largest CBSS area (23.7 $\mathsf{\mu }{\mathrm{m}}^2$) and the thickest peel (13.1 mm); mandarin showed the most directionally oriented flavedo film ($A_F=0.0885$); and lemon showed the most open albedo (${\phi }_{2D}=36.2\%$). Oil-gland equivalent diameter was essentially invariant (∼45 $\mathsf{\mu }$m) across the five fruits, while gland density varied 4.4-fold. The structural metrics define a layered descriptor space—a dense isotropic surface relief versus a thick cellular bulk—that supplies two distinct bioinspired-design priors: dense surface films as a structural prior for selective-permeability membranes and layered cellular cores as a prior for impact-absorbing panels. A modified-atmosphere packaging (MAP)-compatible biomimetic film is identified as one downstream design hypothesis requiring direct gas-permeability verification on synthetic membranes. Full article

The optimized design of key components in harvesting equipment is significantly impeded by the significant grain loss from the header and high energy consumption during stalk cutting that result from the lack of physical and mechanical parameters regarding the plant-flower head system during the mechanized harvesting of edible sunflowers. To furnish the design of mechanized harvesting equipment for palatable sunflowers with theoretical support and foundational data, physical parameters measured included geometrical properties, critical bending angle, coefficient of static friction, moisture content, and head seed collision loss rate. Mechanical parameters—radial elastic modulus, shear modulus, and shear strength—were obtained from stalk compression and shear tests using a universal testing machine. Stem-head detachment force was quantified with a universal testing machine fitted with bespoke fixtures, and orthogonal experiments were conducted with tensile speed, head-picking plate spacing, and tensile angle as factors to establish the significance hierarchy and optimal configuration. Considerable heterogeneity was observed: mean plant height, head diameter, and head thickness were (1733 ± 153) mm, (275 ± 28) mm, and (93 ± 19) mm, respectively. The critical bending angle decreased with height, whereas stalk moisture content increased from base to apex. Mean stalk and head moisture contents were 65% and 61.4%. The coefficient of static friction varied from 0.24 to 0.63 depending on contact material. A critical impact velocity of 2–3 m/s induced mechanical damage and seed cracking. The stalk radial elastic modulus was (1.12 ± 0.27) MPa; shear modulus and shear strength increased with decreasing sampling height, with basal stalks exhibiting a mean shear modulus of 2.47 MPa and shear strength of 1.87 MPa. Sampling position significantly influenced shear modulus (p < 0.05). The factor significance for stem-head detachment force was head-picking plate spacing > tensile angle > tensile speed. The optimal combination (tensile speed 500 mm/min, head-picking plate spacing 50 mm, tensile angle 10°) yielded a detachment force of (202.3 ± 9.5) N, with a relative error below 5% compared to prior detachment force measurements, confirming the reliability of the optimised results. These data provide essential foundations for developing stalk cutting, head inserting, and combine harvesting equipment. Full article

Efficient AGV/AMR path planning is essential for intelligent warehouse logistics, where regular shelves, narrow aisles, local bottlenecks, and heterogeneous obstacles strongly affect roadmap quality. This study proposes LE-HG-PRM, a structure-aware extension of heuristic-guided probabilistic roadmap planning. The method embeds warehouse geometric priors into probability-field sampling, region-adaptive neighborhood connection, and cache-accelerated progressive path refinement. Compared with the preliminary conference version, the journal version introduces a redesigned warehouse-oriented planning framework and substantially expands the experimental validation. Four experimental campaigns are conducted, covering static-complexity progression, corridor-width sensitivity, parameter sensitivity, and map-scale expansion, with A*, JPS, PRM, RRT, RRT*, and HG-PRM as baselines. Each scenario uses 50 paired start–goal tasks, and sampling-based methods are repeated with 12 independent random seeds. The results show that LE-HG-PRM provides competitive path quality and structurally regular paths in representative warehouse layouts. Statistical tests further confirm that its path-length advantage is scenario-dependent but significant in several structured and bottleneck-constrained settings. The findings suggest that incorporating explicit warehouse-structure priors can improve roadmap-based global planning for intelligent logistics, while future work should validate the method in Gazebo and physical AGV/AMR platforms. Full article

Few-shot font generation can be viewed as a challenging conditional image generation task, where the goal is to synthesize target glyphs from only a few reference samples while preserving structural fidelity and style consistency. This problem becomes particularly difficult for characters with complex spatial layouts and fine-grained stroke topology, where existing methods often struggle to simultaneously maintain structural integrity, local continuity, and stylistic coherence under sparse-reference conditions. To address this issue, we propose a skeleton-aware deformable alignment framework for few-shot font generation. Specifically, explicit skeleton priors are introduced into the diffusion-based generation process to provide structural supervision during denoising. In addition, a structure-constrained deformable content alignment module is designed to improve local feature correspondence while suppressing unreasonable geometric deformation. We further develop a multi-module content aggregation strategy to jointly model global layout patterns and local stroke details through complementary multi-level representations. Extensive experiments demonstrate that the proposed method consistently outperforms state-of-the-art approaches in both quantitative and qualitative evaluations. The results show that our method provides stronger structural preservation, better perceptual quality, and improved generalization in structurally complex glyph generation and cross-lingual style transfer. Full article

The objects in high-resolution remote sensing images usually exhibit arbitrary orientations, multi-scale variations, dense distributions, and complex background interference, posing significant challenges to oriented object detection. Although existing DETR-style end-to-end detectors eliminate the need for anchor design and non-maximum suppression, they still suffer from insufficient orientation priors in object queries, limited orientation consistency in decoder feature interaction, and unstable set matching for oriented bounding boxes. To address these issues, this paper proposes an end-to-end Transformer framework, termed OPT-Net (Orientation-Preserving Transformer Network), for oriented object detection in remote sensing images. OPT-Net treats orientation information as a structured geometric prior and propagates it through query initialization, feature interaction, and matching optimization. Specifically, an Orientation-Aware Query Initialization (OAQI) module is designed to generate initial queries using center confidence and orientation priors. An Orientation-Consistent Cross-Attention (OCCA) mechanism is proposed to perform orientation-conditioned modulation on Value features while keeping the standard Query–Key attention computation unchanged. Furthermore, an Uncertainty-aware Matching Loss (UML) is introduced to incorporate instance-level geometric uncertainty into Hungarian matching and regression optimization. Experimental results on the DOTA-v1.0 and HRSC2016 datasets show that OPT-Net achieves 76.83% and 90.58% mAP, respectively, demonstrating competitive detection accuracy and adaptability to complex remote sensing scenarios. Ablation studies and visualization results further validate the effectiveness of each proposed module. Full article

Reconstructing building models from urban SAR tomography (TomoSAR) point clouds is often constrained by limited resolution, low positioning accuracy in elevation, as well as data incompleteness and artifacts caused by microwave imaging mechanisms. These challenges seriously restrict the extraction of high-accuracy building models with structural details from TomoSAR point clouds. This paper proposes a refined urban building modeling method that effectively utilizes structural priors, including directionality, orthogonality, and potential symmetry. First, a piecewise fitting strategy integrated with density-based segmentation is employed to iteratively estimate the main directions of the buildings and capture finer geometric variations of complex façade footprints than simple-plane approximations. Second, a roof extraction algorithm combining an adaptive Doug-las–Peucker approach with symmetry evaluation and constraints is developed to regularize roof outlines and repair data defects. Crucially, to handle extreme cases where roof data are entirely missing, a novel building width estimation method based on building shadow analysis is proposed. Experiments conducted on the SARMV3D-1.0 and SARMV3D-3.0 point cloud datasets demonstrate that the proposed method significantly enhances reconstruction accuracy and geometric fidelity in urban regions compared to state-of-the-art approaches. Full article

Haze commonly degrades visible-spectrum remote sensing (RS) images by reducing contrast and distorting colors. Existing RS dehazing methods still face two limitations. Prior-driven methods rely on handcrafted assumptions that may become unreliable in complex wide-area scenes without explicit sky regions. Learning-based methods require paired training data, yet real aligned hazy/haze-free RS image pairs are difficult to collect, which limits their real-world generalization. To address these limitations, we propose a method called Remote Sensing Image Dehazing via RGB-Space Physical Constraints (RDPC). The new method revisits the atmospheric scattering model (ASM) from the perspective of RS imaging and builds the restoration process on several physical properties of hazy image formation. For atmospheric light estimation, the RGB-space line-convergence behavior of local regions with similar reflectance and slight depth variations is exploited, allowing atmospheric light to be estimated without explicit sky areas. For transmission estimation, the geometric relation between observed pixels and atmospheric light is used in RGB space, where local perpendicularity provides physically plausible haze-removal guidance and global compensation helps avoid excessive darkening and color degradation. The estimated transmission and albedo guidance are further refined by enforcing ASM consistency and variation sparsity through joint optimization. Experiments on synthetic and real-world RS image dehazing benchmarks demonstrate that RDPC achieves competitive performance against representative prior-based and learning-based methods, including Image Dehazing and Exposure (IDE), Iterative Predictor-Critic (IPC), Curvature-to-Plane Prior (C2P), Adaptive Structure-Texture Awareness (ASTA), Asymmetric U-Net (AU-Net), Efficient Multi-scale Prior Fusion (EMPF), and Lightweight Feature Dehazing (LFD), in terms of peak signal-to-noise ratio (PSNR), structural similarity index measure (SSIM), learned perceptual image patch similarity (LPIPS), Blind/Referenceless Image Spatial Quality Evaluator (BRISQUE), neural image assessment (NIMA), and processing time. Full article

Autonomous driving in unstructured environments faces challenges such as missing road boundaries, terrain variations, random obstacle distributions, and complex vehicle–terrain interactions, making it difficult to achieve safe navigation by relying on lane-level priors from structured roads. To address the problems of the relative separation between traversability analysis and trajectory planning, the ineffective propagation of perception uncertainty, and the insufficient scene adaptability of coupling mechanisms, this paper takes traversability as the main thread and systematically reviews the research progress of perception–planning coupling mechanisms in unstructured environments. First, traversability analysis methods based on geometric terrain, semantic understanding, and physical dynamics are reviewed, and the representation and propagation mechanisms of uncertainty in the perception–planning chain are analyzed. Second, the role of traversability information in global path search, local trajectory optimization, and data-driven planning is discussed, and the applicable boundaries of different coupling architectures are summarized from the perspectives of representation level and system organization form. Finally, datasets, simulation platforms, and evaluation metric systems are summarized, and a risk-state-oriented adaptive perception–planning coupling framework is proposed to dynamically adjust coupling strength based on risk-state information, thereby improving the safety, interpretability, and environmental adaptability of autonomous driving in unstructured environments. Full article

Background/Objectives: Left ventricular (LV) geometry reflects structural adaptation to aging and biological sex. While cardiac magnetic resonance (CMR) provides precise morphologic assessment, most prior studies have focused on volumetric and mass-based parameters rather than routinely reported linear indices. This study aimed to evaluate the influence of age and sex on LV geometry using wall thickness, LV end-diastolic diameter (LVEDD), and proportional indices derived from standard CMR reports. Methods: In this retrospective cross-sectional study, 95 adult patients who underwent clinically indicated CMR were included. LV wall thickness, LVEDD, relative wall thickness (RWT), and wall thickness-to- LVEDD ratio (WT/LVEDD) were recorded. Participants were stratified by sex and age groups (18–40, 41–60, >60 years). Group comparisons, correlation analysis, multivariable linear regression, logistic regression, and Age × Sex interaction testing were performed to evaluate independent associated parameters of LV morphology and concentric remodeling. Results: The mean age was 34.94 ± 16.00 years; 60.0% were male. Males had significantly larger LVED (43.12 ± 6.83 mm vs. 39.76 ± 6.11 mm, p = 0.014) and greater wall thickness measurements (p < 0.05 for septal and posterior wall thickness). Age showed a significant positive correlation with mean LV wall thickness (r = 0.275, p = 0.007) and WT/LVEDD ratio (r = 0.241, p = 0.019), but not with LVEDD (p = 0.414). In multivariable analysis, male sex was independently associated with larger LVED (B = 3.345, p = 0.017), whereas age was independently associated with WT/LVEDD ratio (B = 0.0018, p = 0.019). Logistic regression demonstrated that age independently increased the odds of concentric remodeling (OR = 1.041 per year, 95% CI: 1.011–1.072, p = 0.006). No significant Age × Sex interaction was observed. Conclusions: Advancing age was independently associated with proportional LV geometric remodeling, whereas male sex primarily influenced absolute ventricular dimensions. Routine CMR report-derived linear measurements were sufficient to detect these distinct structural patterns. These findings highlighted the feasibility of using standardized morphologic indices in daily clinical practice to identify early age-related concentric remodeling. Full article

Extremely low overlap leads to severely scarce local geometric correspondences across frame pairs. Pure geometric descriptors—encoding merely low-level shape signatures—inherently fail to impose sufficient constraints for reliable transformation estimation when matches become critically sparse, rendering registration fundamentally fragile. While recent red-green-blue-depth (RGB-D) attempts have explored visual augmentation, they predominantly rely on low-level chromatic statistics or shallow convolutional neural network (CNN) features, underutilizing the rich hierarchical semantics inherent in RGB imagery. We present V3Reg, a robust registration framework that pioneers the integration of large-scale vision foundation models (DINOv3) with adaptive cross-modal fusion. Specifically, we extract mid-to-deep semantic features (Layer 11) from DINOv3 to transcend low-level texture limitations, and propose a Task-Aware Channel-Wise Gated Adaptive Fusion (TACGAF) module that dynamically calibrates geometric-visual contributions via registration-error-guided channel-wise gating. To rigorously evaluate ultra-low-overlap robustness, we reconstruct RGBD-ZeroMatch, a benchmark with controllable overlap ratios ranging from 1% to 20%. Extensive experiments demonstrate that V3Reg achieves 99.6% Feature Matching Recall and 96.3% Registration Recall on standard benchmarks. Notably, it maintains 50.2% Registration Recall at merely 5% overlap, outperforming prior methods by over 18 percentage points. Full article

Thermal overload in internal combustion engines may progressively destabilize lubricant-film integrity and promote severe tribological deterioration within highly stressed contact interfaces. This study investigates the thermo-tribological degradation sequence of a high-mileage gasoline engine subjected to prolonged idle operation under impaired cooling conditions, ultimately resulting in engine seizure. The investigated engine had accumulated 356,724 km, while the lubricant had remained in service for approximately 26,724 km prior to the experiment. The post-failure investigation combined teardown inspection, geometrical camshaft assessment, reverse gravimetric reconstruction, hydraulic tappet surface profiling, XRF surface characterization, laboratory oil analysis, and SEM/EDS evaluation of wear debris. The results demonstrated strongly localized degradation concentrated primarily within the cam–tappet interfaces. Severe non-uniform camshaft wear was accompanied by pronounced hydraulic tappet surface damage and evidence of unstable boundary-lubrication conditions. Laboratory oil analysis revealed elevated wear-metal concentrations, depletion of the alkaline reserve, increased oxidation indicators, and a final Class D oil condition assessment. SEM/EDS characterization identified Fe-bearing wear debris associated with sustained material removal and debris recirculation during the final degradation stage. The combined evidence supports a coupled thermo-tribological degradation mechanism involving lubricant deterioration, boundary-lubrication instability, adhesive wear acceleration, oxidative surface degradation, and debris-assisted surface damage preceding final engine seizure. The present case study provides experimentally documented evidence of lubrication collapse under real-engine thermal runaway conditions and highlights the critical role of lubricant condition in maintaining tribological stability under severe thermal loading. Full article