Search Results (1,207)

Underwater image enhancement has been widely studied to improve visual quality; however, its impact on downstream geometric tasks such as sparse 3D reconstruction remains insufficiently understood. In particular, visually enhanced images do not necessarily lead to improved feature matching or reconstruction performance. This work addresses the problem of selecting appropriate preprocessing strategies for underwater Structure-from-Motion (SfM) pipelines from a task-oriented perspective. We propose a lightweight machine-learning-based preprocessing selector that predicts reconstruction performance from image statistics and recommends suitable enhancement strategies for each input sequence. To ensure reliable evaluation, we introduce a grouped leave-one-parent-sequence-out protocol that avoids overlap-induced bias common in clip-wise splitting. Experiments are conducted on challenging underwater datasets derived from the Real-world Underwater Image Enhancement (RUIE) benchmark, with the primary comparison variable defined as the number of reconstructed sparse 3D points. Supporting geometric variables, including the number of registered images, mean track length, and mean reprojection error, are recorded for interpretation. Results show that preprocessing choices significantly affect reconstruction outcomes and that the optimal strategy is scene-dependent. The proposed selector consistently improved over raw input on the evaluated grouped subset and remained competitive with a strong fixed preprocessing baseline. The grouped leave-one-parent-sequence-out protocol is intended to reduce overlap-induced bias common in clip-wise splitting and to provide a more conservative estimate of generalization. This work highlights the importance of task-aware preprocessing and reliable evaluation in underwater vision systems, offering practical insights for deploying enhancement strategies in real-world 3D reconstruction pipelines. Full article

Automotive plastic parts present multiple challenges for Computer-Aided Engineering (CAE) simulation modeling, including complex thin-walled geometries, difficulties in meshing fine features (e.g., clips and snap-fits), and time-consuming manual processing with inconsistent quality. To address these issues, this paper proposes an automated method for generating mid-surface meshes. The proposed approach integrates AI-based feature recognition, point cloud registration, and geometric fitting. First, a specialized point cloud dataset consisting of 132,000 samples of plastic part features was constructed. Using a PointNet++ model, precise semantic segmentation of typical features, such as clips and backing plates, was achieved. Subsequently, a library of typical features was established, and an FPFH-ICP point cloud registration strategy was implemented. Based on the matching rate, an adaptive selection between two processing paths, direct standard mesh replacement and segmentation-fitting generation was performed. For features with low matching rates, a suite of segmentation-fitting algorithms was proposed. These algorithms incorporate incomplete cylinder parameter extraction, Monte Carlo boundary identification, and internal point cloud reordering, thereby facilitating high-quality mid-surface mesh generation for complex topological structures. Finally, experimental validation was conducted on typical automotive interior plastic parts as well as on new cross-platform vehicle models. The results demonstrate that the proposed method reduces mesh modeling time by 67% while preserving the accuracy of geometric feature restoration. The mesh quality compliance rate increases from 52.27% to 90.9% with the proposed method, reaching a level comparable to that of professional manual meshing. In cross-platform validation, the proposed method maintained high accuracy. Consequently, this approach significantly enhances the intelligence and engineering reliability of CAE pre-processing, providing effective technical support for the automated simulation modeling of complex thin-walled components. Full article

During seafloor terrain mapping missions conducted by AUVs, positioning error accumulation occurs inevitably over long distances due to the unavailability of global satellite navigation signals underwater. Moreover, the alternating distribution of flat and undulating regions on the seafloor renders single-constraint-based bathymetric SLAM methods prone to performance degradation in complex environments. To address these challenges, this paper proposes a multi-constraint hybrid strategy SLAM framework for AUV-based seafloor terrain mapping, grounded in an analysis of error accumulation mechanisms and constraint failure characteristics. The framework establishes a hierarchical and progressive constraint architecture to enable collaborative optimization across different spatial scales and topographic conditions. At the foundational pose estimation stage, multi-source trajectory information is fused to ensure continuity and stability in pose computation. In the local consistency constraint stage, an improved point cloud registration method combined with a neighborhood survey-line constraint mechanism is introduced to enhance geometric consistency among survey lines in feature-sparse regions. At the global optimization stage, a loop closure detection strategy is designed based on topographic statistical features, incorporating adaptive thresholds and correlation metrics to achieve robust introduction of global constraints. By flexibly integrating direct registration and feature-matching strategies according to topographic characteristics, the framework fully leverages the advantages of multi-constraint cooperative optimization. The proposed method is validated by the field data. Experimental results on real lake-trial data show that, relative to the baseline configurations evaluated under identical noise-injection conditions, the MCHS-SLAM framework yields more concentrated consistency-error distributions with markedly shorter large-error tails, and exhibits improved error suppression relative to the reference trajectory. This work presents a methodological framework for high-quality seafloor terrain mapping under heterogeneous terrain conditions, providing a basis for future extensions toward onboard real-time deployment. Full article

Indoor environments often contain numerous areas with sparse structural features, such as long corridors, large atriums, and glass curtain walls, and other scenarios. These conditions can lead to difficulties in loop closure detection and accumulated positioning errors, resulting in localization drift or even mapping failure during map construction. This paper proposes an indoor mapping algorithm called IS-LEGO-LOAM that integrates tightly coupled LiDAR-IMU fusion and Scan Context. A tightly coupled LiDAR-IMU odometry is constructed, and an adaptive covariance matrix is designed to solve the problems of abnormal LiDAR echoes and insufficient effective feature extraction caused by sparse indoor feature points. By introducing the Scan Context global descriptor and adopting the strategies of vector nearest neighbor search and similarity score matching, the drift problem in large-scale scenes is alleviated. Finally, validation is performed on the KITTI dataset and in real-world scenarios, respectively. Experiments show that the improved IS-LEGO-LOAM achieves superior mapping performance. Full article

Accurate and robust UAV navigation in complex urban environments remains challenging due to dense buildings, dynamic obstacles, and unreliable GPS signals. To address this issue, this paper proposes GLP-VO, a hybrid visual odometry framework that combines geometric structure features with point features. An adaptive weighting strategy is introduced to balance the contributions of different feature types according to matching quality and scene complexity, while geometric constraints are incorporated into the optimization process to improve pose estimation accuracy and stability. Experiments on the TUM RGB-D dataset and real UAV flight sequences verify the effectiveness of the proposed method. GLP-VO achieves the best ATE results in five of the ten evaluated TUM sequences, including 0.91 cm on f1_xyz and 0.62 cm on f3_str_tex_far, and remains competitive on challenging sequences such as f2_360_kidnap with an ATE of 2.26 cm. In the ablation study, the full model reduces ATE and RPE by up to 44.9% and 43.1%, respectively. Moreover, the proposed system runs at approximately 35 FPS on the desktop platform and 11 FPS on the onboard platform, demonstrating a favorable balance between accuracy, robustness, and real-time performance. Full article

Rapid urbanization and the rapid accumulation of multi-source and multi-temporal Earth observation data are creating an increasing demand for remote sensing models that can flexibly support fine-grained monitoring beyond fixed label taxonomies. Open-vocabulary remote sensing image semantic segmentation (OVRSIS) aims to segment text-specified categories beyond a fixed label space with vision–language foundation models. However, dense remote sensing scenes make pixel–text matching highly vulnerable to semantic confusion and misalignment, owing to extreme scale variation, thin structures, repetitive textures, and prompt sensitivity. To address these challenges, we propose UGDMoE, an uncertainty-guided mixture-of-experts framework for OVRSIS. First, we design a domain-specific MoE decoder with three geometrically specialized experts—for slender structures, mid-scale objects, and large-region context—routed by the alignment-risk cue $U_0$. Second, we introduce a lightweight prompt–response estimation strategy that quantifies prediction dispersion across semantically equivalent prompts to derive $U_0$ in an annotation-free manner. Third, we develop prompt ensemble-based likelihood calibration (PELC), which takes the shared alignment-risk cue $U_0$ as input to calibrate prompt-specific logits before refinement. Finally, we design a lightweight uncertainty-aware structure refinement module that, guided by $U_0$, selectively fuses early visual features with segmentation logits to restore boundary continuity and connectivity of thin structures. We conduct extensive experiments on eight OVRSIS benchmarks under cross-dataset evaluation protocols. Trained on DLRSD, it achieves 46.97 m-mIoU and 63.31 m-mACC, surpassing the strongest baseline by 0.76 and 0.62 points; trained on iSAID, it reaches 37.47 m-mIoU and 58.52 m-mACC, improving over the strongest competitor by 0.71 and 0.61 points. UGDMoE consistently achieves state-of-the-art performance and remains robust under training-source changes. Full article

Many place recognition approaches, which identify previously visited places or locations by matching current sensory data, such as 2D RGB images and 3D point clouds, have been proposed to achieve accurate and robust localization and loop closure detection in global positioning system (GPS)-denied environments. Since visual place recognition (VPR) methods that rely on images captured by camera sensors are highly sensitive to variations in appearance, including changes in lighting, surface color, and shadows, they can lead to poor place recognition accuracy. In contrast, light detection and ranging (LiDAR)-based place recognition (LPR) approaches based on 3D point cloud data that captures the shape and geometric structure of the environment are robust to changes in place appearance and can therefore provide more reliable place recognition results than VPR methods. This work presents an indoor LPR method called PointNetVLAD-based indoor pedestrian localization (PIPL). PIPL is a deep network model that uses PointNetVLAD to learn to extract global descriptors from 3D LiDAR point cloud data. PIPL can recognize places previously visited by a pedestrian using point clouds captured by a low-cost LiDAR sensor on a smartphone in small-scale indoor environments, while PointNetVLAD performs place recognition for vehicles using high-cost LiDAR, GPS, and inertial measurement unit (IMU) sensors in large-scale outdoor areas. For place recognition on 3D point cloud reference maps generated from LiDAR scans, PointNetVLAD exploits the universal transverse mercator (UTM) coordinate system based on GPS and IMU measurements, whereas PIPL uses a virtual coordinate system designed in this study due to the unavailability of GPS indoors. In experiments conducted in campus buildings, PIPL shows significant advantages over NetVLAD (known as a convolutional neural network (CNN)-based VPR method). Particularly in indoor environments with repetitive scenes where geometric structures are preserved and image-based appearance features are sparse or unclear, PIPL achieved $39\%$ higher top-1 accuracy and $10\%$ higher top-3 accuracy compared to NetVLAD. Furthermore, PIPL achieved place recognition accuracy comparable to NetVLAD even with a small number of points in a 3D point cloud and outperformed NetVLAD even with a smaller model training dataset. The experimental results also indicate that PIPL requires over $76\%$ less place retrieval time than NetVLAD while maintaining robust place classification performance. Full article

Existing technologies can achieve relative geometric correction and stabilization of geostationary satellite image sequences through fixed land scene matching or homonymous point adjustment. However, these methods heavily rely on fixed land areas, rendering them completely ineffective in vast ocean regions with only ship targets. Additionally, the trajectories of ship targets after processing still exhibit noticeable jitter, hindering motion information analysis. To address these issues, this paper proposes a joint image adjustment and stabilization method based on multi-target trajectories in marine environments: (1) An optimized target detection algorithm based on a multi-scale heterogeneous convolution module is introduced, which extracts background and target features through convolutions of different scales, enabling accurate detection and tracking of weak small targets in the image sequence frame by frame. (2) Curve fitting is performed on the detected positions of the same ship across multiple frames to simulate its motion trajectory under stabilized conditions. Combined with the prior assumption of uniform motion, an equal-division strategy is adopted to determine the corrected positions of the target in the image sequence. (3) The deviation correction values of multiple targets within the same frame are obtained, and based on the principle of intra-frame deviation consistency, precise image stabilization is achieved under multi-target constraints. Experiments based on Gaofen-4 satellite image sequences demonstrate that this method reduces the average position deviation of ship targets in the original images from 8.5 pixels (425 m) to 3.4 pixels (170 m), a decrease of approximately 59.41%, effectively improving the relative geometric accuracy of the image sequence and significantly eliminating target trajectory jitter. Full article

Accurately identifying urban functional zones and revealing their spatial interaction characteristics is crucial for understanding urban operational mechanisms and optimizing spatial layouts. Addressing the limitations of traditional research in simultaneously capturing static functional attributes and dynamic resident travel behaviors, this study takes the central urban area of Lhasa as the research object, integrating ride-hailing trajectory data with Point of Interest (POI) data to conduct research on urban functional zone identification and spatial interaction characteristics. First, Thiessen polygons were used to quantify the spatial influence range of POIs, and an address matching algorithm was employed to associate ride-hailing origins and destinations (ODs) with POIs. A weighted land use intensity index was constructed, and functional zones were precisely identified using information entropy and K-Means clustering. Secondly, with basic research units as nodes and OD flows as edges, a directed weighted spatial interaction network was constructed. Complex-network indicators and the Infomap community detection algorithm were utilized to analyze network characteristics, node importance, and community interaction patterns. The results show that: (1) The functional mixing degree in the study area exhibits a pattern of “highly composite core, relatively differentiated periphery.” Eight functional zone types, including commercial–residential mixed, science–education–culture, and transportation service zones, were ultimately identified. Residential areas form the base, while the core area features multi-functional agglomeration. (2) The spatial interaction network exhibits typical small-world effects, while its degree distribution is better characterized by a lognormal distribution rather than a power law. Node importance is dominated by betweenness centrality, with Lhasa Station, the Potala Palace, and core commercial areas constituting key hubs. (3) The network can be divided into four functionally coupled communities: the core multi-functional area, the western industry–residence integrated area, the eastern science–education-dominated area, and the southern transportation hub area, forming a “core leading, two wings supporting” center–subcenter spatial organization pattern. This study verifies the effectiveness of integrating trajectory and POI data for identifying urban functional zones and provides a new perspective for understanding the spatial structure and planning of plateau cities. Full article

Wind turbine blades are subjected to complex environmental conditions during long-term operation, which may lead to structural degradation and performance loss. To ensure structural integrity, fatigue testing prior to deployment is essential. This paper proposes a vision-based method for measuring the full-cycle breathing deformation of wind turbine blades during fatigue testing. The method captures dynamic image sequences of the blade’s hotspot cross-section using industrial cameras and employs a feature-based template matching approach to reconstruct the three-dimensional coordinates of target points. Through coordinate transformation, the deformation trajectories are obtained, enabling quantitative analysis of the blade’s dynamic responses in both flapwise and edgewise directions. A dedicated hardware–software system was developed and validated through full-scale fatigue experiments. Quantitative comparison with strain gage measurements shows that the proposed method achieves mean absolute deviations of 0.84 mm and 0.93 mm in two independent experiments, respectively, with closely matched deformation trends under typical loading conditions. These results demonstrate that the proposed method can reliably capture the global deformation behavior of the blade with millimeter-level accuracy, while significantly reducing instrumentation complexity compared to conventional contact-based approaches. The proposed method provides an effective and practical solution for full-field dynamic deformation measurement in blade fatigue testing, offering strong potential for structural health monitoring and early damage detection in wind turbine systems. Full article

Currently, China boasts abundant shale gas resources. However, in the process of flowing production, there remain significant discrepancies in our understanding of the flow patterns of gas and water, and many challenges persist in gas–water measurement. Given the dense pore structure and complex micro-features of shale gas reservoirs, this study proposes a method to estimate the fractal dimension by utilizing shale mercury injection curves based on experimentally determined relative permeability curves, thereby enabling a more accurate fitting of these curves. Experimental results show that the two-phase co-infiltration zone in the shale is narrow overall, with bound water saturation exceeding 50%. The findings indicate that the experimentally measured relative permeability curves closely match those fitted using the fractal dimension approach. Moreover, the lower the permeability, the more the equal-permeability points of the fitted curves shift toward the lower-right quadrant. Overall, the fitting performance is satisfactory, providing additional research directions and insights for determining relative permeability curves of gas and water in shale gas reservoirs. Full article

Artificial intelligence (AI), cloud computing, robotics, automation, and remote sensing technologies are all contributing to digital transformation in forestry. Improving on low-accuracy Global Navigation Satellite Systems (GNSS) positioning affected by multipath error and interception under forest canopies is critical for integrating smart and digital technologies into equipment in forest operations. In an era where lidar-derived individual tree locations are now increasingly available in digital forest inventories, a possible alternative approach to positioning resources such as people or equipment accurately could be to match locally-measured tree positions and attributes in the forest with an existing global reference map based on prior remote sensing missions, effectively using the trees themselves as satellites to circumvent the need for GNSS-based positioning. We evaluated a lidar-based alternative to GNSS positioning using predicted tree positions from local terrestrial laser scanning (TLS) matched with a global stem map derived from prior airborne laser scanning (ALS), a methodology we refer to as TreePS. The horizontal error of the TreePS system was estimated using 154 permanent single-tree inventory plots on the University of Idaho Experimental Forest with two different workflows based on two common R packages (lidR v. 4.3.0, FORTLS v. 1.6.2) using either spatial coordinates or spatial plus stem DBH predicted using one or both segmentation routines and a custom matching algorithm. Mean TreePS error using lidR for below and above-canopy segmentation had mean error of 1.04 and 2.04 m with 93.5% and 91.6% of plots with viable match solutions on spatial and spatial plus DBH matching. The second workflow with both FORTLS (TLS point cloud) and lidR (ALS point cloud) had errors of 1.09 and 2.67 m but only 57.9% and 54.2% of plots with solutions using spatial and spatial plus DBH, respectively. There is room for improvement in the matching algorithm but the TreePS methodology and similar feature-matching solutions may be useful for below-canopy positioning of equipment, people or other resources under dense forests and other GNSS-degraded environments to help advance smart and digital forestry. Full article

Low-orbit thermal infrared bidirectional whisk-broom imaging offers wide-swath coverage and high spatial resolution for monitoring moving targets such as aircraft, but large scan angles and terrain undulation cause non-rigid geometric distortion and radiometric inconsistency between forward and backward scans. These effects generate strong clutter in difference images and degrade small and weak target detection. To address this problem, we propose an infrared small target detection method that fuses accurate registration and weighted difference. First, we propose a hybrid multi-scale registration algorithm that achieves coarse affine registration through sparse feature–point matching and then iteratively corrects nonlinear deformations by integrating a global grayscale-driven force with a local sparse-feature-guided force, yielding a registration error of 0.3281 pixels. On this basis, a multi-scale weighted convolutional morphological difference algorithm is proposed. A novel dual-structure hollow top-hat transform is constructed to accurately estimate the background, and a multi-directional convolution mechanism is introduced to effectively suppress anisotropic edge clutter and enhance target saliency. Experiments on SDGSAT-1 thermal infrared bidirectional whisk-broom data show an SCRG of 18.27, and a detection rate of 91.2% when the false alarm rate is below 0.15%. The method outperforms representative competing algorithms and provides a useful reference for space-based aerial moving target detection. Full article

Odometry estimation, which calculates the trajectory of a moving object across timeframes, is a critical and time-consuming function in SLAM (Simultaneous Localization and Mapping) systems. Although LiDAR-based sensing is most popular for outdoor and long-range applications because of its ranging accuracy, the sparsity of laser point cloud poses a significant challenge to feature extraction and matching in odometry estimation. In this paper, we investigate odometry estimation from two aspects, i.e., algorithm optimization, and system design/implementation. In algorithm optimization, we present an image feature-assisted odometry estimation scheme that leverages the richness of image information captured by a companion camera to enhance the accuracy of laser point cloud matching. This also serves as a screening mechanism to reduce the matching size and lower the computing complexity for a higher estimation rate. In addition, various schemes, such as adaptive threshold in image feature point selection, principal component analysis (PCA)-based plane fitting for laser point interpolation, and Gauss–Newton optimization for calculating the transform matrix, are also employed to improve the accuracy of odometry estimation. The performance of improved odometry estimation is verified using an existing FLOAM (Fast Lidar Odometry and Mapping) framework. The KITTI dataset for autonomous vehicles with ground truth was used as the test bench. Simulation results indicate that the translation error and rotation error can be reduced by 16.6% and 1.3%, respectively. Computing complexity, measured as the software execution time, also reduced by 63%. In system implementation, a hardware/software (HW/SW) co-design strategy was adopted, where complexity profiling was first conducted to determine the task partitioning and time-consuming tasks are offloaded to a hardware accelerator. This facilitates real-time execution on a resource-constrained embedded platform consisting of a microprocessor module (Raspberry Pi) and an attached FPGA board (Pynq Z2). Efficient hardware designs for customized DSP functions (adaptive threshold and PCA) were developed in an FPGA capable of completing one data frame in 20ms. The final system implementation met the target throughput of 10 estimations per second, and can be scaled up further. Full article

LiDAR point cloud semantic segmentation is essential for autonomous driving, yet LiDAR-only methods remain constrained by sparsity and limited texture cues. We propose Cross-Modal Collaborative Manifold Distillation (CMCMD), which transfers open-world semantic priors from the DINOv3 Vision Foundation Model to a LiDAR student network. The framework combines an Adaptive Relation Convolution (ARConv) backbone with geometry-conditioned aggregation, a Unified Bidirectional Mapping Module (UBMM) for explicit 2D–3D interaction, and Manifold-Aware Topological Distillation (MATD), which aligns inter-sample affinity structures in a shared latent manifold rather than enforcing pointwise feature matching. By preserving relational topology instead of absolute feature coordinates, CMCMD mitigates negative transfer across heterogeneous modalities. Experiments on SemanticKITTI and nuScenes yield mIoU values of 72.9% and 81.2%, respectively, surpassing the compared distillation baselines and approaching the performance of multimodal fusion methods at lower inference cost. Additional evaluation on real-world campus scenes further supports the cross-domain robustness of the proposed framework. Full article