Search Results (571)

The accuracy of shallow-water multibeam bathymetry is critically dependent on precise heave correction. However, sensor limitations often lead to incomplete correction, leaving periodic along-track stripe noises (heave residuals) that distort seabed morphology. Traditional filtering methods suppress this noise at the expense of genuine topographic detail. This paper proposes an innovative Data Feedback Correction (DFC) method that corrects the error at its source. DFC establishes a closed-loop framework: it diagnoses the residual’s dominant frequency from central beam data, extracts the residual signal via targeted filtering, and feeds it back as a compensation term into the original sensor heave sequence. This drives a recomputation of the geometric positioning, achieving source-level correction. In a field case, DFC demonstrated targeted, high-fidelity performance. Across all 34 survey lines, DFC achieved an average spectral attenuation of 1.85 dB (range: 1.0–3.7 dB) in the dominant residual band and reduced the RMSE of overlap discrepancies from 0.0923 m to 0.0773 m (a 16.25% improvement). Independent validation using 94,999 control line intersections further demonstrates a 14.31% RMSE improvement relative to an uncorrected control line reference, confirming that the correction improves both internal consistency and external accuracy, significantly enhancing internal consistency. Compared to moving average and wavelet denoising, DFC achieved comparable quantitative improvement while effectively suppressing visual stripes and features that are consistent with the original data, avoiding the over-smoothing or residual noise of traditional methods. This study confirms that closed-loop feedback of data residuals can fundamentally address spectrally aliased stripe noise, shifting the paradigm from “masking noise” to “correcting the source.” The method enhances data consistency in the tested scenario without sacrificing topographic authenticity, providing a promising new tool that warrants further validation across diverse survey conditions. Full article

Modern fulfillment centers share floor space with human workers, making warehouse multi-robot task allocation a safety-critical problem. We propose SHARP (Safe Heterogeneous Allocation with Risk Prediction), a Transformer-based constrained reinforcement-learning framework with a closed-form deployment-time safety mask. Under a Gaussian pedestrian belief and fixed closest-approach directions, the mask uses Bonferroni-allocated per-pair CVaR scores; a nonnegative mask score implies a conservative trajectory-level chance constraint under the stated assumptions. We also present an idealized primal–dual surrogate analysis, without claiming global convergence for the nonconvex Transformer/PPO implementation. Expanded experiments use ten training seeds per learned method and deterministic final-checkpoint evaluation on twenty independently generated held-out instances. No statistically significant difference between SHARP and Lagrangian-PPO was detected in any of the four scenarios. The held-out analysis further reveals late-training instability and severe over-conservatism in the dense S40_high scenario. These findings position SHARP as an auditable geometric filtering mechanism, while identifying conservatism and training stability as important limitations for deployment. Full article

Phosphate waste rock piles (PWRPs) generated by open-pit phosphate mining are highly heterogeneous and difficult to characterize using conventional point sampling alone, which limits representative resource assessment, selective recovery, and rehabilitation planning. This study develops an integrated framework combining PRISMA spaceborne hyperspectral imagery, ground-based mineralogical analyses, and spatially constrained machine learning to map lithological heterogeneity at the Benguerir phosphate mining site, Morocco. A three-stage spectral optimization workflow, including atmospheric band masking, Savitzky–Golay filtering, and analysis of variance (ANOVA)-based feature selection, was applied to identify the most discriminative Short-Wave Infrared (SWIR) bands for lithological classification. After removing redundant observations located within shared PRISMA pixel footprints, 127 spatially independent samples were retained for model development. Five supervised classifiers (Random Forest, Extra Trees, XGBoost, Support Vector Machine, and K-Nearest Neighbors) were evaluated under a spatially constrained cross-validation framework aligned with the 30 m native PRISMA pixel size. Ensemble-based models, especially Extra Trees and Random Forest, provided the most stable performance, with balanced accuracies of 0.56–0.69 and area under the receiver operating characteristic curve (AUC) values exceeding 0.95 for carbonate-dominated lithologies. Lower discrimination between phosphate and siliceous facies reflects intrinsic mineralogical mixing and spectral overlap at the sensor scale. Entropy-based uncertainty and posterior probability mapping revealed spatially structured prediction ambiguity concentrated along lithological boundaries and transitional zones, consistent with petrographic evidence of compositional heterogeneity. These results indicate that moderate but stable accuracies likely represent realistic performance limits for spaceborne hyperspectral mapping of complex mining environments under spatial constraints. The proposed framework provides a transferable and uncertainty-aware basis for lithological mapping, selective recovery assessment, and sustainable phosphate waste management. Full article

In the application of unmanned aerial vehicle (UAV) target perception in complex marine environments, the significant cyclostationarity of UAV radar echoes makes it highly suitable for extracting their signatures via cyclic spectral analysis. This method projects the signal onto the cyclic frequency dimension, exploiting the fundamental difference between the periodicity of the UAV’s micro-vibrations and the non-periodic randomness of sea clutter, enabling the effective and reliable extraction of the UAV’s target features. However, the sea-clutter background often masks the UAV signal, making it difficult to identify the target processing unit for cyclic spectral analysis rapidly. Autocorrelation processing excels at rapidly filtering out non-periodic components from the echo signal, thereby preserving and enhancing periodic components. It exploits the correlation between adjacent pulses to suppress slow clutter and enhance the echoes from moving targets, thereby establishing a target range for cyclic spectral analysis. Inspired by this, we first propose a novel method in this paper that innovatively employs autocorrelation-guided cyclic spectral fusion filtering, which effectively mitigates the short-term coherence and non-stationarity characteristics of strong sea-clutter background. Corresponding results with a measured strong sea-clutter background demonstrate that the proposed method effectively suppresses sea clutter and reliably extracts UAV target signals from other maritime targets. Compared with the classic moving target indicator (MTI) and the singular value decomposition (SVD) method, as well as their cascade processing, the proposed method achieves higher gain across various input signal-to-clutter-plus-noise ratios (SCNRs), demonstrating broad applicability and excellent detection performance. Full article

This study evaluates Retuve, an open-source explainable pipeline for the automated analysis of infant hip ultrasound cine sweeps. Retuve combines segmentation, Graf-plane calibration, and frame filtering. In a retrospective multicenter study, we tested the full pipeline on an external set of 109 hips from a Canadian community clinic, with internal developmental validation of segmentation on 90 hips and Graf-plane calibration on 419 hips. On the external test set, Retuve achieved 100% specificity and 91% sensitivity for expert agreement regarding whether a sweep contained an analyzable frame, compared with 75% specificity and 96% sensitivity for a radiology fellow; specificity was based on 16 expert-negative examinations. For alpha angle and acetabular coverage, Retuve achieved consistency intraclass correlation coefficients (ICCs) of 0.77 and 0.74, comparable to the fellow’s 0.70 and 0.74. However, alpha-angle absolute agreement was lower (ICC 0.55, 95% confidence interval (CI) −0.07–0.81), consistent with systematic measurement bias. Internal developmental validation showed Component 1 mask mean average precision at 50% intersection-over-union (mAP50) of 0.753 and box mAP50 of 0.883 and a Component 2 ICC of 0.792. Retuve can select analyzable frames and recover measurements from variable-quality cine sweeps, but alpha-angle calibration requires refinement. Future prospective work should evaluate developmental dysplasia of the hip (DDH) diagnostic accuracy, clinical treatment decision support, and screening outcomes. Full article

Traditional visual SLAM systems often suffer from localization drift in dynamic environments due to interference from moving objects. Although semantic segmentation and depth-based masking methods have improved performance, they may still suffer from boundary under-segmentation and missed detections due to truncation of dynamic objects. To address these challenges, we propose a cascaded framework, DMSG-SLAM, a cascaded visual SLAM system that fuses Depth-Mask, Semantic information and Geometry constraints for dynamic environments. A lightweight object detection network, combined with depth consistency, is first employed to generate instance-like masks for preliminary dynamic feature removal. Then, a rotation-aware local epipolar geometric filtering mechanism is introduced to suppress residual features near object boundaries and mitigate perceptual blind spots caused by occlusion or truncation. Within potential dynamic regions, the epipolar threshold is adaptively switched according to the estimated inter-frame rotation to provide a more conservative filtering effect under challenging motion conditions. In addition, a TSDF-based dense volumetric map is incorporated to reconstruct more consistent surfaces. Experiments on highly dynamic sequences from the TUM RGB-D dataset indicate that DMSG-SLAM achieves competitive accuracy in dynamic environments, with localization performance improving by up to 90% compared to ORB-SLAM2. Full article

Semantic segmentation of large-scale airborne point clouds traditionally relies on labor-intensive 3D manual annotations. While recent zero-shot methods attempt to alleviate this burden by distilling knowledge from 2D Vision–Language Models (VLMs) via 2D-to-3D projection, they suffer from performance degradation in complex urban environments. Specifically, lacking 3D geometric awareness, 2D VLMs frequently exhibit “semantic bleeding”, where large-scale background categories (e.g., ground) erroneously submerge small-scale targets (e.g., vehicles and street elements). To address this issue, we propose a geometry-constrained pseudo-label generation and purification framework. Our approach tackles the problem through a dual-branch design: extracting open-vocabulary semantics via SAM3-based multi-view projection while simultaneously deriving sharp, class-agnostic instances using SAM2 on Gamma-transformed elevation maps. By introducing a geometric–semantic consistency module, we evaluate the internal semantic purity and external spatial homogeneity of these instances, detecting and filtering out semantic misclassifications. The purified pseudo-labels are then used to supervise a 3D sparse convolutional network via a Masked Cross-Entropy Loss. Experiments on the H3D and Turin3D datasets demonstrate that our method recovers small-scale targets that are prone to being submerged, outperforming existing zero-shot baselines by improving mIoU from 52.15% to 63.45% on H3D and from 29.52% to 58.51% on Turin3D, thereby narrowing the performance gap with fully-supervised approaches. Full article

Group equivariant convolutional neural networks (G-CNNs) achieve superior sample efficiency by encoding symmetry into network architecture, yet their computational overhead (up to 3.78× slower inference and 4.63× more multiply–accumulate operations) hinders deployment on resource-constrained edge devices. Existing pruning methods cannot be applied directly: arbitrarily removing weights breaks the group representation structure and degrades equivariance. We characterize the complete design space of equivariance-preserving compression, proving that exactly two axes leave a convolutional layer equivariant: irrep-bundle pruning, which reduces irreducible-representation multiplicities, and orbit-wise pruning, which removes complete spatial orbits from kernel supports; via Schur’s lemma, no third structure-preserving axis exists. This completeness result, rather than the use of representation theory itself, is our central contribution. We turn it into practice through direct sub-filter extraction, which yields real convolutional parameter reduction (up to 83%) and 1.4–2.9× measured inference speedup, unlike masking, which gives no real speedup. Across three datasets (MNIST, CIFAR-10, EuroSAT) and three symmetry groups (C₄, D₄, SO(2)), compression is nearly lossless on strongly symmetric data: the 4-layer EuroSAT model drops only 1.07% at 83% reduction. On weakly symmetric data (CIFAR-10), the pruned model can even gain 2.6 points, but our analysis attributes this to relaxing a mismatched equivariance constraint rather than to pruning itself; the value of pruning over from-scratch training scales with the data’s symmetry strength. Full article

Recovering high-quality images from low-quality speckle patterns remains a core challenge in scattering imaging, especially under narrowband illumination with ambient light interference and broadband illumination. This paper proposes a dual-input synchronous transmission architecture: after Correction-Smoothing-Phase Optimization (CSPO) preprocessing, two data streams are parallel-fed into Phase-Aligned Coherent Summation (PACS) for efficient and high-precision reconstruction with adaptive fusion, breaking the single-path limitation of traditional methods and balancing imaging efficiency and quality. Additionally, an adaptive enhancement factor feedback mechanism is designed for Median-Unsharp Sharpening Enhancement (MUSE) to dynamically adjust Median Filtering (MF) and Unsharp Masking (USM) parameters, achieving adaptive balance between noise suppression and detail enhancement and improving robustness under extreme lighting. In PACS, a dynamic reference update mechanism is introduced, combined with fixed amplitude to realize iterative phase optimization, effectively suppressing speckle noise and boosting the signal-to-noise ratio of reconstructed images. Experimental results show that the proposed method achieves favorable restoration performance even at a SNR of −8.7 dB under narrowband and broadband illumination with spectral bandwidths of 100 nm, 200 nm, and 280 nm (FWHM), and significantly improves image quality in unknown scattering media, showing great potential for robust speckle reconstruction. Full article

Wireless sensor networks (WSNs) deployed for seismic monitoring must sustain long-term operation under strict energy constraints, where premature node failure degrades spatial coverage and detection reliability. This paper presents a safety-constrained reinforcement learning framework for transmission scheduling in energy-harvesting seismic WSNs. The proposed approach integrates Proximal Policy Optimisation (PPO) with action masking and a runtime guard-layer safety filter that enforces battery-preservation and load-balancing constraints without retraining. The guard layer intercepts policy actions and substitutes safe alternatives when constraint violations are detected, using a scoring function that combines battery headroom with network-wide load equity. Experiments across three network scales (10, 15, and 30 nodes) with solar energy harvesting demonstrate that the guard-enhanced PPO achieves 99.46% transmission success at 30 nodes while maintaining 66.47% node survival—a 58.3% improvement in survival over the highest-reward baseline (Closest) at the cost of only a 6.2% reduction in cumulative reward. Crucially, the guard-enhanced policy outperforms the unconstrained PPO baseline simultaneously on cumulative reward ($+11.4\%$), transmission success ($+0.8$ pp), and node survival ($+15.4\%$), demonstrating that hard safety constraints, when properly aligned with the system’s energy model, provide both performance and safety gains rather than a fundamental trade-off. Sensitivity analysis across event rates ($p_{\mathrm{event}}=0.5$ and $0.9$) confirms that the guard layer’s advantage persists under both moderate and extreme monitoring conditions. Analysis across scales reveals distinct operational regimes: at 10 nodes, heuristic baselines are near-optimal; at 30 nodes, learned policies dominate, and safety filtering becomes critical for sustained operation. Full article

Real-time kinematic (RTK) positioning in open-pit mining environments is critically compromised by non-line-of-sight (NLOS) signals and anisotropic multipath effects induced by pit walls, haul roads, and industrial infrastructure. Conventional elevation-dependent stochastic models fail to discriminate between geometrically favorable low-elevation satellites and those subject to directional obstruction, resulting in degraded ambiguity resolution and decimeter-level positioning errors that undermine safety-critical deformation monitoring. This paper presents an adaptive RTK positioning framework utilizing azimuth-integer-based elevation masking to explicitly model site-specific obstruction geometry. The proposed method discretizes the horizontal plane into 360 integer-degree sectors, extracts minimum elevation angles per sector from 24 h line-of-sight (LOS) data, and constructs a smoothed 360°mask profile via moving-window filtering. A virtual elevation-angle transformation is introduced to normalize satellite geometry relative to the local mask, enabling adaptive down-weighting of diffraction-susceptible observations within the stochastic model without requiring multi-day satellite repeat arcs or hardware modifications. The approach was validated using 54 h of BDS data collected at eight monitoring stations within the Wangjialing open-pit mine, China. Implementation of the mask model engendered a selective 8.1% reduction in satellite participation (15.66 to 14.39 satellites) while significantly enhancing observation quality. The ambiguity validation ratio improved by 19.5% (from 9.43 to 11.27 in the experimental project), and the fix success rate increased from 92.4% to 97.2% (exceeding the 95% reliability threshold at all stations). The RMS errors in the east, north, and up directions improved by 34.8% to 65.2%, 28.7% to 77.0%, and 44.8% to 70.8%, respectively, with the most dramatic gains observed at stations subject to severe azimuthal obstruction (e.g., ZDH6 vertical RMS: from 50.7 mm to 14.8 mm). By explicitly modeling anisotropic obstruction geometry through discrete angular sampling, the proposed method achieves sub-centimeter positioning accuracy and robust ambiguity resolution in challenging mining environments without additional hardware or empirical threshold tuning, offering a cost-effective solution for large-scale, real-time deformation monitoring systems. Full article

Aiming at the challenges of class imbalance, tiny defect scales, and complex brushed background interference in the surface defect detection of electrical commutators, this paper proposes a high-precision and lightweight improved instance segmentation algorithm named WG-YOLOv11. Firstly, to overcome the barrier of highly imbalanced positive and negative samples in actual industrial data collection, a Balanced Defect Synthesis (BDS) data augmentation strategy is introduced to effectively enrich the morphological diversity of tiny defects. Secondly, a Wavelet Transform Convolution (WTConv) module is collaboratively integrated into the feature extraction network to expand the receptive field while preserving the high-frequency edge details of hairline cracks. Thirdly, a Group CBAM Enhancer (GCE) module is introduced to filter out high-reflection and brushed background noise through grouped attention and weight re-calibration mechanisms. Finally, addressing the difficulty of pixel-level alignment for tiny defects, an $\alpha$-IoU loss function is utilized to improve the high-precision segmentation and localization capabilities by dynamically adjusting the gradient distribution. Comprehensive evaluations are conducted on two real-world electrical commutator surface defect datasets: KolektorSDD2 and KolektorSDD. Experimental results show that on the KolektorSDD2 dataset, compared to the YOLOv11 baseline, the Mask mAP@50 of WG-YOLOv11 increases from 85.2% to 89.2%, and the stringent metric Mask mAP@50:95 improves from 52.7% to 56.9%. Additional computational analysis on the same dataset validates that the proposed method maintains high efficiency, matching the baseline computational cost without compromising real-time inference speed. Furthermore, evaluations on the public MSD dataset confirm the model’s cross-domain generalization capabilities. The proposed framework effectively achieves a balance between detection accuracy, anti-interference robustness, and a lightweight architecture. Full article

The widespread manipulation of digital images on social media has significantly undermined public trust in visual content and created major challenges for automated forgery detection. These challenges are further intensified by platform-induced degradations such as compression, resizing, and filtering, which often obscure forensic traces. This work develops FIDD-6000, a large-scale benchmark dataset for image forgery detection and localization, containing 6000 social media images, including 1000 authentic and 5000 manipulated samples, with pixel-level ground-truth masks annotated across three forgery categories, splicing, copy-move, and retouching, all created under realistic post-processing conditions. Each manipulated image is accompanied by a pixel-level ground-truth mask indicating the tampered regions. To assess the challenges posed by social media-based image manipulation, we evaluate 15 state-of-the-art image forgery localization methods on FIDD-6000, including approaches based on JPEG compression artifacts, sensor-noise analysis, and error level analysis. Experimental results show that these methods perform poorly on the proposed dataset, revealing their limited effectiveness in detecting forged images that have undergone social media-specific compression and transformation. This performance gap highlights the need for more robust and advanced machine learning and deep learning approaches capable of handling the complexity of modern image manipulations. Therefore, FIDD-6000 provides a valuable resource for researchers by offering a rigorous benchmark for developing, evaluating, and comparing next-generation forgery detection and localization methods. Full article

To address the severe performance degradation of radar weak target detection induced by dense cluster targets and sea-spike interference in nonhomogeneous sea clutter environments, this paper proposes an enhanced Adaptive Normalized Matched Filter algorithm based on iterative weighting and sample purification (IWP-ANMF). The proposed algorithm establishes a closed-loop iterative detection framework capable of highly sensitive discrimination of anomalous data within the reference window—particularly cluster targets and strong discrete sea spikes that severely distort covariance matrix features—identifying them as “contaminated samples.” During each iteration, target-likelihood statistics are calculated for all reference samples based on the current covariance matrix estimate. Subsequently, an adaptive deep-notch suppression strategy is applied to contaminated samples, such as cluster targets, according to their statistical characteristics, thereby progressively purifying the sample covariance matrix (SCM) estimation. Theoretically, this iterative procedure is rigorously proven to converge to the optimal solution of a robust weighted covariance matrix estimation problem. Comprehensive validations using both Monte Carlo simulations and measured K-distributed sea clutter data demonstrate that, compared to classical ANMF and Generalized Inner Product (GIP) approaches, the proposed algorithm exhibits outstanding robustness and detection performance when confronted with heterogeneous contamination scenarios, especially high-density cluster targets. This method effectively eliminates the blind-zone expansion and performance deterioration caused by the wideband masking of cluster targets, significantly enhancing weak target detection capabilities under complex maritime conditions. Full article

Accurate target speech mask estimation is the key to single-channel speech separation. Masks generated by conventional mask networks are easily corrupted by interfering speech and background noise, which degrades separation performance. To solve this problem, this paper proposes a Cross-Purification Mask Network (CPMN), which consists of three core modules: the Dynamic Context-Aware Mechanism (DCAM), Feature Cross-Complementation Mechanism (FCCM), and Adaptive Purification Mask Mechanism (APMM). The DCAM aggregates dynamic sliding window and long-term temporal features to capture long-range temporal dependencies of masks and enhance the localization accuracy of target speech. The FCCM fuses weighted mask features of interfering speakers to dynamically supplement missing information in target speech masks. The APMM combines adaptive filters and residual networks to output high-precision refined masks. The CPMN is embedded into three mainstream speech separation frameworks including Conv-TasNet, DPTNet, and TDANet, and extensive experiments are conducted on Libri2Mix, WHAM!, and WSJ0-2Mix datasets. The results show that the CPMN brings stable performance gains. After integration, TDANet achieves SI-SNRi of 17.4 dB (+0.5 dB) on Libri2Mix and 15.2 dB (+0.4 dB) on WHAM!. Meanwhile, Conv-TasNet and DPTNet obtain SI-SNR improvements of 0.3 dB (15.6 dB) and 0.4 dB (20.8 dB) on WSJ0-2Mix, respectively. Full article