TerraMosaic Daily Digest: July 13, 2026
Daily Summary
A frictional criterion for the transition from intermittent creep to catastrophic landslide failure provides the clearest mechanistic advance in today's literature. Ring-shear experiments and long-term monitoring identify a metastable regime of pore-pressure-driven slip pulses, progressive dilation and diminishing rate strengthening before sustained acceleration. Complementary studies move from this failure threshold to forecasting and consequence: evolutionary feature selection reduces rainfall-induced landslide prediction error in the Southern Andes, while field, three-dimensional and rheological analyses resolve loess-slide reactivation, fault-controlled instability, rockfall runout and landslide loading of pipelines. Multisource inversion further constrains concealed water-sand inrush pathways where direct subsurface imaging is infeasible.
Hazard observation is becoming both denser and more explicitly probabilistic. A 450-km ocean-bottom fibre-optic array images a distant intermediate-depth earthquake at kilometre scale; ensemble tsunami forecasts propagate stochastic slip uncertainty into inundation maps; foundation-model fusion converts pixel-level change masks into building-level damage counts without instance annotations; and flood reconstructions in Ukraine and the Danube separate persistent upper-level forcing, catchment saturation and tributary synchronisation. For urban infrastructure, models of ground-fissure crossings, seismic slope tests and spatially variable soil-structure interaction quantify where geometric assumptions or unresolved heterogeneity alter damage estimates.
Environmental context is treated increasingly as a dynamic boundary condition rather than a static covariate. Interpretable machine learning maps gully density across the Qinghai-Tibet Plateau; climate ensembles project a 53-day expansion of its mean non-frozen season by 2100; a daily 1-km dataset resolves six atmospheric-moisture indicators across China; and long-term water-storage reconstruction extends the GRACE record back to 1961. Saltmarsh and restored-mangrove studies independently show how canopy structure and seasonality regulate storm-wave attenuation. Together with new geospatial reasoning, foundation-model and multimodal sensing benchmarks, these results tighten the connection between observation design, physical process and decision-relevant hazard estimates.
Key Trends
Five methodological shifts connect today's hazard, observation, and modelling studies.
- Failure thresholds replace descriptive precursors: Intermittent slip pulses, dilation and frictional state evolution are organized into a testable instability criterion for slow-moving landslides.
- Forecasting shifts toward compact, uncertainty-aware predictors: Genetic feature selection, stochastic tsunami ensembles and spatially variable seismic models expose which inputs control skill and where confidence should narrow.
- Sensing geometry becomes part of the inference: Ocean-bottom DAS, drone photogrammetry, InSAR, LiDAR and hydrogeophysics are evaluated together with their spatial coverage, transfer limits and failure modes.
- Coupled mechanics and hidden pathways link hazards to damage: Fault geometry, soil arching, tunnel-fissure intersection and inferred water-sand pathways are incorporated directly into slope, pipeline and underground-risk models.
- Hydroclimate memory is resolved across timescales: Flood antecedence, terrestrial water storage, freeze-thaw seasonality and vegetated wave damping show how delayed or seasonal forcing reorganizes hazard intensity.
Selected Papers
The leading studies replace empirical warning signals with testable process constraints. Ring-shear experiments identify a metastable pulse-and-dilation regime before catastrophic landslide acceleration; Southern Andes forecasting isolates precipitation, slope and soil-hydraulic controls; and stochastic tsunami ensembles propagate slip uncertainty into inundation. Field-to-model analyses then show how fault geometry, loess rheology, vegetation, soil arching and concealed water-sand pathways govern failure and consequences.
1. Precursors of catastrophic failure in intermittent creep landslides: a frictional evolution perspective
Core Problem: Slow-moving landslides can alternate between creep and transient acceleration without a physical criterion that identifies when this behaviour becomes unstable.
Key Innovation: Ring-shear experiments and long-term monitoring reveal a metastable regime of pore-pressure-driven slip pulses and dilation; a critical friction threshold marks the transition to sustained acceleration and catastrophic failure.
2. High-Resolution Imaging of a Distant M6 Intermediate-Depth Earthquake Using Ocean-Bottom DAS and Seismic Network
Core Problem: Intermediate-depth earthquakes are difficult to image at regional distance with the spatial resolution needed to distinguish rupture sub-events.
Key Innovation: A 450-km seafloor DAS array and conventional network resolve two sub-events and 6-8 km of northward propagation in about three seconds for an Mw 6 earthquake more than 400 km away.
3. AI-based ensemble tsunami inundation forecasting
Core Problem: Near-field tsunami forecasts must deliver inundation rapidly while representing uncertainty in earthquake slip rather than relying on one uniform-slip scenario.
Key Innovation: An AI ensemble trained on 630 scenarios combines Green's-function waveforms with stochastic slip models, raising median inundation accuracy to 92% and producing an explicit probabilistic forecast.
4. Feature selection for landslide forecasting models in Southern Andes
Core Problem: Rainfall-induced landslide warning in the Southern Andes is constrained by heterogeneous inventories, 136 candidate variables and no standard way to identify a compact predictor set.
Key Innovation: A 3,148-instance database, balanced sampling and genetic feature selection reduce model error to 10.95% with random forest and XGBoost while repeatedly identifying precipitation, slope and soil-water storage as dominant controls.
5. Mechanism of the Huangci 2# landslide reactivation in December 2025 and discussion of its causes (Yanguoxia Town, Gansu, China)
Core Problem: The December 2025 reactivation of the 4.5-million-cubic-metre Huangci 2# landslide requires attribution across irrigation, rainfall, stratigraphy and pre-existing creep.
Key Innovation: Field mapping, remote sensing, drilling, monitoring and rainfall records reconstruct the reactivation mechanism and document burial of 43 houses, farmland and critical local infrastructure.
6. Three-dimensional stability analysis of fault-controlled slopes using a column-based composite slip-surface framework
Core Problem: Conventional three-dimensional stability methods idealize slip geometry and can miss failure surfaces constrained by faults.
Key Innovation: A composite fault-rock-mass surface, residual-thrust formulation and differential-evolution search reproduce the observed open-pit failure with a factor of safety of 1.013.
7. Synergistic effects of water content and temperature on the rheological behavior of loess and its implications for loess geohazards
Core Problem: Loess-flow mobility depends on coupled water and temperature effects that cannot be represented by a monotonic viscosity correction.
Key Innovation: Rheometry reveals water-dependent thermal thinning, aggregation and a non-monotonic transition regime; the Herschel-Bulkley model captures these responses better than the Bingham model.
8. Vegetation and Geomorphological Controls on Rockfall Hazard Along Volcanic Slopes
Core Problem: Rockfall runout on the urbanized volcanic slopes of Campi Flegrei depends on discontinuity geometry, channelization and vegetation that are rarely evaluated together.
Key Innovation: Drone-derived terrain, structural mapping and RAMMS simulations show threefold longer runout on channelized slopes and substantial forest attenuation, although blocks can still reach the urban area with energies near 6,000 kJ.
9. Mechanical Response and Soil Arching Mechanism of Buried Pipelines in Landslide Zones
Core Problem: Pipeline design in landslide zones lacks a quantitative treatment of three-dimensional soil arching and its transition across moving and stable ground.
Key Innovation: A validated mechanical model partitions tension and compression zones and shows how burial depth, landslide width and friction govern axial stress, displacement and abrupt changes in vertical arching.
10. Seismic behavior of metro shield tunnels traversing ground fissure zones: Deformation characteristics and quantitative damage assessment
Core Problem: Tectonic ground fissures impose differential settlement and localized seismic demand on metro tunnels, but damage varies with both shaking level and crossing angle.
Key Innovation: A validated three-dimensional soil-fissure-tunnel model quantifies lining damage and identifies a 60-degree intersection as comparatively favourable while mapping angle-dependent damage zones.
11. TerraLogic: A Benchmark for Hierarchical Geospatial Reasoning in Earth Observation
Core Problem: Earth-observation benchmarks test perception more often than the hierarchical reasoning required for hazard vulnerability and multi-step geospatial analysis.
Key Innovation: TerraLogic contributes 545 multimodal tasks and an open tool-augmented baseline that organizes functions hierarchically, recovers from tool failures and exposes large remaining gaps in geospatial reasoning.
12. The catastrophic floods in 2008, 2010 and 2020 in western Ukraine: Hydrometeorological processes and the role of upper-level dynamics
Core Problem: The 2008, 2010 and 2020 floods in western Ukraine arose from different combinations of atmospheric persistence, moisture transport and antecedent wetness.
Key Innovation: Case analysis and a 22-event climatology attribute 64% of heavy-precipitation events to potential-vorticity streamers and distinguish the forcing pathways of the three catastrophic floods.
13. The September 2024 Danube flood compared to the 1899, 2002, and 2013 events: a hydrometeorological analysis in a changing climate
Core Problem: The September 2024 Danube flood cannot be interpreted from basin-average rainfall because impacts were concentrated in saturated tributaries under persistent forcing.
Key Innovation: Station, reanalysis and gauge data show 450-500 mm local rainfall totals, runoff coefficients of 0.6-0.9 and a shift toward regionally concentrated, convectively enhanced flood generation.
14. Seismic response of the slope under different boundary conditions through shaking table test
Core Problem: Shaking-table estimates of slope response can be dominated by container boundaries, particularly as excitation amplitude and frequency increase.
Key Innovation: A damping-fluid boundary treatment reproduces the enlarged-model response under low-amplitude, low-frequency shaking and defines when larger physical models remain necessary.
15. Insta-BDA: Instance-Aware Building Damage Assessment and Counting via Foundation Model Fusion
Core Problem: Pixel-level post-disaster change maps do not preserve individual buildings, limiting operational damage counts and per-structure assessment.
Key Innovation: Insta-BDA fuses ChangeMamba with zero-shot SAM3 instances using only pixel-level damage labels; on xBD it improves aggregate building-count deviation from -47.06% to -34.96%, raises damage F1 from 0.67 to 0.70 and generalizes better across two independent datasets.
16. Identifying equivalent migration pathways for water-sand inrush in coal mines by integrating multisource data and sensitivity analysis
Core Problem: Water-sand inrush can propagate through concealed, combined subsurface pathways that cannot be imaged directly before or after failure.
Key Innovation: Multisource sediment and groundwater evidence, numerical inversion and global sensitivity analysis identify a confined aquifer at 707-724 m depth and place the most likely pathway inlets 310-390 m from the mine inrush site.
17. Methodology for assessing earthquake-induced road interruption due to structural and nonstructural debris applied to a virtual testbed
Core Problem: Earthquake debris from both structural collapse and nonstructural damage can block roads and delay response, but city-scale interruption estimates rarely represent both sources.
Key Innovation: A census-section methodology estimates road-blockage probability from building density, typology, age and street layout and demonstrates the framework on representative Italian urban forms.
18. Effect of burial depth on the seismic dynamic response of rectangular subway station considering soil spatial variability
Core Problem: Deterministic soil profiles understate how burial depth and spatial variability jointly alter the seismic response of underground stations.
Key Innovation: Random-field simulations show increasing response dispersion and failure risk with depth, with many samples collapsing at 20 m burial under the tested earthquake loading.
19. Semantic-Edge Response Decoding of SAM3 for Zero-Shot Crack Segmentation
Core Problem: Zero-shot foundation-model masks suppress or over-expand the thin, fragmented evidence needed for crack inspection, while supervised alternatives require pixel-level labels.
Key Innovation: SERD decodes SAM3's internal semantic responses as a dense crack field and calibrates them with an edge prior; without annotation or fine-tuning, it reaches 61.14% mean Crack IoU across six datasets, 4.63 points above native SAM3.
20. Self in Space: Benchmarking Self-Awareness and Spatial Cognition in UAV Embodied Intelligence
Core Problem: UAV benchmarks emphasize the surrounding scene but rarely test whether an embodied model tracks its own motion, memory and position within that scene.
Key Innovation: SIS-Bench contributes 4,856 expert-verified questions across 13 tasks and 1,646 real UAV videos; motion-aware optical-flow fusion improves both self-awareness and spatial cognition and transfers to downstream decision tasks.
21. Label-Decoupled Style Augmentation for Domain Generalization in Multi-Label Remote Sensing Scene Classification
Core Problem: Global feature-statistics augmentation can mix the styles of unrelated labels within a multi-label aerial scene, contaminating cross-domain training.
Key Innovation: Label-specific attention confines style mixing to regions that share a class; on a three-domain benchmark, the best variant reaches 71.5% mean average precision, five points above empirical risk minimization, with at most 0.35% additional parameters.
22. An Extension to the Procedure for Developing Uncertainty-Consistent Shear Wave Velocity Profiles from Inversion of Experimental Surface Wave Dispersion Data
Core Problem: Uncertainty-consistent shear-wave velocity profiling could not previously combine surface-wave arrays of different sizes into one broadband, deep site model.
Key Innovation: Two reconstructions of the full dispersion-data matrix recover inter-wavelength correlation and uncertainty-consistent velocity suites on synthetic sites, while derived Vs30 and site-period proxies remain substantially better constrained than individual profiles.
23. MBTI: A Multi-Branch Efficient Fine-Tuning Framework for Hyperspectral Image Classification with Foundation Models
Core Problem: Foundation-model transfer between hyperspectral sensors often compresses or selects bands to fit a fixed input, discarding information and breaking local spectral continuity.
Key Innovation: MBTI partitions the full spectrum into continuous branches, adapts each with LoRA and fuses them by channel attention; it is competitive across three datasets while training only 2.33-2.36% of model parameters.
24. Dispersion-Guided Physics-Aware Deep Inverse Operator for Surface Wave Mode Separation
Core Problem: Overlapping fundamental and higher surface-wave modes make dispersion-curve picking and shear-wave velocity inversion unreliable, especially in two-station analysis.
Key Innovation: A label-free inverse operator uses adaptive frequency-phase-velocity masks as physical constraints to separate modes directly in time-space data; synthetic and field tests improve automated dispersion picking and subsequent inversion.
25. ViCo3D: Empowering LiDAR-based Collaborative 3D Object Detection with Vision Foundation Models
Core Problem: Collaborative LiDAR perception lacks the semantic priors of vision foundation models, but direct transfer is impeded by the image-point-cloud modality gap.
Key Innovation: ViCo3D renders point clouds as three-channel bird's-eye-view images for DINOv2, then fuses visual and geometric features across agents; it achieves state-of-the-art detection and up to 1.8 times the collaborative gain of prior methods on DAIR-V2X.
26. The Emerging Paradigm of Geospatial Foundation Models: From Pre-Training to Agentic Reasoning
Core Problem: Geospatial foundation models differ in pretraining, adaptation cost and operational constraints, leaving no consistent route from pretrained model to domain deployment.
Key Innovation: The review separates self-supervised vision and vision-language GeoFMs, organizes adaptation strategies by cost and capability, and extends the framework to tool-mediated agentic geospatial reasoning.
27. ABot-N1: Toward a General Visual Language Navigation Foundation Model
Core Problem: Monolithic visual-language navigation policies accumulate coordinate drift, handle rare semantics poorly and obscure how high-level intent becomes motion.
Key Innovation: ABot-N1 separates a slow vision-language reasoner that emits pixel goals from a fast waypoint controller; POI arrival rises by 35 points to 77.3%, with success rates of 95.4% indoors and 92.9% outdoors, and new navigation benchmarks are released.
28. Wave attenuation through a saltmarsh: Heterogeneous vegetation characteristics and seasonal variability
Core Problem: Saltmarsh wave models depend on an empirical dissipation coefficient that changes with vegetation structure, submergence, season and wave processing choices.
Key Innovation: A momentum-coupled spectral formulation derives dissipation from measurable canopy properties and reproduces laboratory and Tropical Storm Lee observations without prescribing the coefficient.
29. Benefits of the simplified MEV for analyzing hourly precipitation extremes in a changing climate
Core Problem: Hourly precipitation return levels are highly uncertain for short records, and the behaviour of competing extreme-value models changes under non-stationary warming.
Key Innovation: Convection-permitting climate simulations show that sMEV is less sensitive than GEV to small samples, that their return-level ordering reverses with warming, and that both tested non-stationary forms remain too simple for unqualified projection.
30. HiMIC-Daily: A high-resolution (daily and 1 km) multi-indicator atmospheric moisture collection over China, 2003–2020
Core Problem: China lacks a daily, kilometre-scale atmospheric-moisture product that resolves multiple indicators across complex terrain and short-lived hydroclimatic extremes.
Key Innovation: HiMIC-Daily combines 2,419 stations, ERA5-Land, land-surface temperature, topography and seasonality with LightGBM to map six indicators at 1 km for 2003-2020; validation R2 ranges from 0.877 to 0.989 and the data are openly available.
31. Drought dynamics across the hydrological cycle - an extensive validation of the National Hydrological Model of Denmark
Core Problem: Drought propagation from precipitation through soil moisture, streamflow and groundwater is rarely tested against long, compartment-specific observations at national scale.
Key Innovation: Up to 34 years of Danish observations show that the national model reproduces groundwater and streamflow anomalies and timing well, with median correlations of 0.76 and 0.79, but exposes materially weaker soil-moisture skill.
32. Nucleation Characterization of Rock Fracture Based on Unsupervised Machine Learning
Core Problem: Fractures nucleate inside loaded rock and cannot be reconstructed reliably from acoustic-emission points without separating noise, microcrack clusters and intersecting planes.
Key Innovation: DBSCAN denoising, expectation-maximization clustering and confidence ellipsoids convert acoustic-emission events into continuous fracture geometries, recovering their locations and orientations and enabling time-resolved nucleation analysis.
33. Deterioration Behavior and Damage Mechanism of High-Stress Anchored Rock Mass Under Cyclic Impact Disturbance Loads
Core Problem: Anchored rock masses under repeated dynamic disturbance can fail abruptly, yet the observable progression from damage accumulation to ejection remains poorly constrained.
Key Innovation: Cyclic impact tests combine digital image correlation, acoustic emission, thermography and discrete elements, linking b-value decline and energy release to failure while resolving how bolts suppress crack propagation.
34. Spatial modeling of gully density on the Qinghai-Tibet plateau: application of hyperparameter optimization in interpretable machine learning
Core Problem: Large-area gully-density prediction remains less accurate than binary susceptibility mapping and lacks a consistent account of controlling factors.
Key Innovation: Bayesian-optimized XGBoost trained on 14,187 one-square-kilometre quadrats reaches 73.5% class accuracy, while SHAP identifies slope, elevation and vegetation cover as dominant controls.
35. Future reshaping of soil freeze-thaw seasonality on the Tibetan plateau
Core Problem: Warming alters not only the duration but also the seasonal timing and temperature thresholds of Tibetan Plateau soil freeze-thaw transitions.
Key Innovation: Bias-corrected CMIP6 ensembles project freezing 30.6 days later, thaw completion 41.4 days earlier and a mean 52.9-day expansion of the non-frozen season by 2100.
36. Reconstruction and evolution analysis of long-term terrestrial water storage anomalies in Xinjiang based on time series decomposition and multi-model coupling
Core Problem: The short and discontinuous GRACE record limits diagnosis of multi-decadal water-storage change in arid Xinjiang.
Key Innovation: Time-series decomposition, VIC and machine learning reconstruct monthly 0.25-degree terrestrial water storage from 1961 to 2022 and reveal declines across 60.8% of the region.
37. Wave damping ability of restored mangrove wetlands during typhoons: A case study from Hailing Island, China
Core Problem: The storm-wave protection delivered by restored mangroves remains poorly constrained relative to mature forests and design costs.
Key Innovation: Thirteen days of Typhoon Maliksi observations show 39% wave-height reduction across 100 m of restored mangrove; a vegetation-drag model identifies diminishing marginal benefit and a site-specific optimum near 16% of total transect width.
38. Undifferenced Integer Ambiguity Resolution in GNSS Network Solutions: Benefits to Satellite Orbits, ERP, Geocenter, and Station Coordinates
Core Problem: Practically, however, several studies showthatthelatterpreserves better results, which isnotclearly explained yet.Inthis study, weinvestigate theperformance ofserval commonly usedDD-IAR andUD-IAR strategies across key GNSS products, including satellite orbits, station coordinates, earthrotation parameters (ERP), andgeocenter coordinates.
Key Innovation: Theoretically, DD-IAR andUD-IAR should achieve equivalent results.
39. Physics-Aware AI-Driven 4D Temperature Reconstruction With Enhanced Non-Uniform Vertical Resolution in Near Space (20-80 km)
Core Problem: Near‐space temperature is crucial for atmospheric dynamics and aerospace applications, yet data sets with refined vertical resolution remain scarce.
Key Innovation: The authors present the Atmospheric Spatio‐Temporal Reconstruction Network (AtmoST‐Net), which integrates high‐resolution SABER observations with dual‐source reanalysis, combining non‐uniform vertical enhancement and explicit physical forcing factors (as input features) to reconstruct temperature at 20–80 km, improving the mid‐to‐upper atmosphere vertical resolution by 3.6 km.
40. Measuring Sand and Finer Suspended Sediment Concentration Profiles in Gravel Bed Rivers
Core Problem: New empirical equations for the scaling factor β in the Rouse number formulation were developed using both total and skin‐friction shear velocities, achieving low uncertainty (relative RMSE < 5%).
Key Innovation: The methodological framework established in this study provides a practical and transferable approach for measuring vertical suspended sediment distributions in rivers with diverse flow and sediment conditions, offering valuable data for improving sediment transport modeling.
41. ACZ-GSeg: Adaptive Concentric Zone-based Two-stage Ground Segmentation for LiDAR Point Clouds
Core Problem: In the coarse segmentation stage, a lowest-height seed constraint and height-decay weighting are introduced to establish a weighted principal component analysis plane fitting model, from which ground candidate points are extracted.
Key Innovation: Experimental results show that the proposed method achieves Precision, Recall, and F1-score values of 99.12%, 96.24%, and 97.66% on the SemanticKITTI dataset, and 98.72%, 100.00%, and 99.36%, respectively, on a self-collected point cloud acquired using a RUBY-PLUS.
42. From Reconstruction to Interpretation: Zero-Setup Multi-Phase Segmentation of X-ray Tomography Data
Core Problem: However, rapid interpretation remains limited by image segmentation, which often requires manual thresholding, user prompting, or material-specific model training.
Key Innovation: The authors present a zero-setup framework for multi-phase segmentation of synchrotron X-ray tomography data that generates interpretable masks for previously unseen datasets without user input or retraining during deployment.
43. MobileSAM2: Lightweight Segment Anything for Spatial Intelligence
Core Problem: However, many of such use cases require to operate on resource-constrained devices like mobile phones and laptops.
Key Innovation: To this end, we propose Hypergraphical Knowledge Distill (HyperKD), which introduces the idea of hypergraph into knowledge distillation, aiming to effectively model and transfer SAM2's generalizable and comprehensive knowledge.
44. DM-KG: A Novel Method for Boosting Spatial Cognition of Vision-Language Models in Street View Imagery
Core Problem: However, existing VLMs frequently suffer from "spatial semantic hallucinations" when perceiving object locations, distances, and directions in real-world street view scenes.
Key Innovation: By establishing a complete, augmented reasoning pipeline, this research significantly improves the spatial cognitive capabilities of VLMs in street view scenarios, thereby providing a flexible, generalized, and interpretable framework for geographic visual question answering (GeoVQA) in open environments.
45. DeGuNet: Depth-Guided Ultra-Compact Backbones for Efficient LiDAR-Camera 3D Detection
Core Problem: However, current multi-modal frameworks heavily rely on massive visual backbones pretrained on 2D semantic tasks.
Key Innovation: When integrated into established baselines, it fundamentally eliminates architectural redundancy, reducing GPU memory consumption by up to 66.5% and achieving a 1.16x inference speedup.
46. More Than Where You Are: Learning Semantics, Structure, and Geometry from Cross-View Localization
Core Problem: Consistent cross-view understanding under extreme viewpoint changes is essential for spatial intelligence, as it enables models to recognize the same scene across extreme viewpoint gaps.
Key Innovation: In this work, we revisit cross-view localization as more than pose estimation and investigate how it can help the model develop consistent cross-view understanding under extreme viewpoint changes, including stable semantics, reliable structure, and transferable geometry.
47. Edge-Aware Thermal Infrared UAV Swarm Tracking
Core Problem: However, tracking tiny UAVs remains challenging due to limited appearance cues, frequent occlusions, and rapid maneuvers.
Key Innovation: More sophisticated nonlinear estimators can improve robustness but often introduce additional computational costs.
48. Domain-Incremental Remote Sensing Change Detection via Difference-Guided Adaptation and Frequency-Decoupled Distillation
Core Problem: Existing domain-incremental learning (DIL) methods mainly preserve image-level representations but often overlook bitemporal discrepancy cues, which are critical for robust change detection under domain shifts.
Key Innovation: To address this limitation, we propose DG-FDD, a domain-incremental change detection framework that integrates Difference-Guided Adaptation and Frequency-Decoupled Distillation.
49. Image Matching Filtering and Refinement by Planes and Beyond
Core Problem: Moreover, a novel and effective strategy combining non-deep traditional computer vision approaches based on planar constraints and cross correlation is presented.
Key Innovation: Unlike previous comparisons, the designed benchmark also takes into account the more general, real, and practical cases where camera intrinsics are unavailable.
50. 2.5D Transformer: An Efficient 3D Seismic Interpolation Method without Full 3D Training
Core Problem: However, its core operation introduces heavy computational burden due to the quadratic complexity, hindering its further application to higher-dimensional data.
Key Innovation: To achieve Transformer-based three-dimensional (3D) seismic interpolation, we propose a 2.5-dimensional Transformer network (T-2.5D) that adopts a cross-dimensional transfer learning (TL) strategy, so as to adapt the 2D Transformer encoders to 3D seismic data.
51. FoundationGeo: Learning Spatial Pixel-Wise Fields for Monocular Metric Geometry
Core Problem: Beyond model design, we identify camera intrinsic coverage, especially focal length distribution mismatch between training and test data, as a key bottleneck for zero-shot metric generalization: performance drops sharply when test intrinsics fall outside the training distribution.
Key Innovation: The authors present FoundationGeo, a two-stage framework that explicitly bridges relative and metric prediction via spatial calibration and principled data design.
52. ABot-3DWorld 0: A Universal World Model to Explore Any 3D Space
Core Problem: The authors present ABot-3DWorld 0, a universal multimodal 3D world model that turns text, image, and video inputs into high-fidelity, explorable 3D worlds.
Key Innovation: The result is one low-barrier engine for general 3D content creation that further anchors generated worlds to geographic points of interest, enabling map-native spatial exploration at consumer scale.
53. Cyclic Loading-Induced Volume Change in Angular and Rounded Gravel-Sand Mixtures in Centrifuge Models
Core Problem: Seismic settlement in gravel-dominated soils is poorly constrained because particle shape, sand migration and gravel-skeleton density interact under cyclic loading.
Key Innovation: Four centrifuge models shaken at about 0.37 g show that, at comparable gravel-skeleton density, rounded mixtures generate greater excess pore pressure and reconsolidation settlement through upward flow and sand migration.
54. Dynamic responses of floating offshore wind turbines under freak waves: A review
Core Problem: Freak-wave loading of floating offshore wind turbines couples transient hydrodynamics with platform, turbine and mooring responses, yet evidence on modelling, experiments and failure modes remains fragmented.
Key Innovation: The review connects freak-wave characterization and reconstruction with numerical and experimental response methods, structural failure pathways and emerging AI prediction approaches, and identifies the principal unresolved modelling needs.
55. Analytical model for the migration and infill of channels in coastal environments
Core Problem: The study presents a new Analytical model for the Migration and Infill of channels in Coastal Environments (AMICE).
Key Innovation: This simple analytical model can be used to explore the morphological behavior of channels, trenches, pits and nourishments in coastal environments, thereby supporting the planning, design and maintenance of these human interventions.
56. Weak-to-Gold Label Training of Spatio-Temporal Deep Learning Models for Cross-Region Coffee and Forest Segmentation
Core Problem: However, there are limited previous works on automated coffee classification utilizing deep learning methodologies, particularly with regards to separating it from forest, as well as across multiple different regions.
Key Innovation: In order to overcome this limitation, we created a dataset containing a large number of poor-quality “weak” labels and a very small number of high-quality gold-standard hand-annotated labels, associated with time series remote sensing patches.
57. HyperMODE: A Continuous-Depth Spectral-Spatial Modeling Framework With Mamba and Neural Ordinary Differential Equations for Hyperspectral Image Classification
Core Problem: However, existing convolutional, Transformer-based, and recent state-space architectures mainly rely on fixed discrete layerwise transformations, which may lead to fragmented intermediate feature updates and insufficient coordination between multiscale local representation learning and global spatial propagation.
Key Innovation: To address this issue, we propose HyperMODE, a unified continuous-discrete spectral-spatial modeling framework for HSI classification.
58. KG-Mamba: A Structure-Aware and Knowledge-Guided Mamba Network for Hyperspectral Image Classification
Core Problem: However, existing Mamba-based methods suffer from three fundamental limitations: loss of spatial continuity from flattening-based sequence modeling, weak representation of local spatial and spectral details, and a semantic gap between extracted features and real-world land-cover classes.
Key Innovation: To address these challenges, we propose knowledge graph (KG)-Mamba, a novel HSI classification architecture that, for the first time, integrates KG-guided semantic projection with structure-aware state-space modeling.
59. Cross-Domain Few-Shot Infrared Ship Segmentation With Class-Specific Adapters and SAM Refinement
Core Problem: However, the significant domain gap between IR and visible images, combined with the extreme scarcity of annotated samples in the target IR domain, poses a dual challenge that severely restricts the cross-domain application of segmentation networks.
Key Innovation: Experimental results on the VI-Ship and Agriculture-Vision datasets demonstrate that the proposed method outperforms state-of-the-art approaches across multiple evaluation metrics.
60. Gridless GLRT Approach for Polarimetric SAR Tomography
Core Problem: Yet, its performance is often limited by the discrete height grid used in conventional reconstruction methods, leading to off-grid errors and biased estimates.
Key Innovation: Tomographic synthetic aperture radar is a fundamental imaging technique for urban environments' mapping and monitoring.
61. MSM_RoadNet: Multispectral Deep Network for Road Extraction With Application to Temporal Change Detection
Core Problem: However, this field still faces three major challenges: the lag in multispectral data supply (with most applications still relying on single-source RGB imagery), the insufficient fusion of multispectral features and topology modeling in traditional models, and the disconnection between segmentation and detection functionalities.
Key Innovation: The work constructed a dedicated multispectral road dataset (MCRData) based on Sentinel-2 imagery to compensate for the limited spectral representation of traditional RGB data.
62. Hyperspectral and Multispectral Image Fusion via Double Decomposition Network
Core Problem: Although detail injection in high-frequency multispectral images (MSI) can improve the spatial resolution of hyperspectral image (HSI) reconstruction, existing methods generally focus on local feature enhancement and lack effective constraints on the detail injection process and global spatial-spectral relationship.
Key Innovation: To address this issue, we propose a double decomposition network (DD-Net), which integrates intrinsic decomposition, a dual-branch structure, and directional enhancement.
63. A long-term dataset on hydrology and suspended sediments in the Kamech catchment from the OMERE Observatory
Core Problem: Small Mediterranean catchments lack long, high-frequency records that resolve rainfall, runoff and suspended-sediment transport across nested spatial scales.
Key Innovation: The Kamech observatory releases quality-controlled records from four nested stations spanning 1.3 to 263 ha, with the longest series approaching 30 years and documented acquisition and processing procedures.
64. Version 3.0.2 of the Crocus snowpack model
Core Problem: It allows the quantification of simulations uncertainty for various applications.
Key Innovation: This article presents a comprehensive description of the 3.0.2 stable release of the Crocus snowpack model in the SURFEX modelling platform.
65. Beyond Runoff Coefficient: Revealing Global Patterns of Process Connectivity in Runoff Generation through Intensity Integration
Core Problem: It is critical to understand this connectivity for climate change adaptation and water-related risk management.
Key Innovation: Here we develop a novel framework to assess process connectivity in runoff generation through intensity integration.
66. Mechanical deterioration and crack coalescence of double-fissured marl under vibration disturbance
Core Problem: Digital image correlation was used to characterize full-field deformation and damage evolution.ResultsThe results show that vibration disturbance markedly reduced the peak strength of marl, with strength deterioration ranging from 6.97% to 23.18%.
Key Innovation: A damage deterioration model incorporating vibration duration and fissure inclination was established, with a coefficient of determination of (R² = 0.89), indicating an acceptable fitting performance in describing the coupled damage evolution of double-fissured marl within the tested parameter range.DiscussionThese findings provide useful insights into the stability assessment of fissured rock masses and the prevention of vibration-induced geological hazards in reservoir areas.
67. Fed-RSAdapter: Federated Fine-Tuning of Remote Sensing Images via Multi-Scale Adapter Modules
Core Problem: In edge computing scenarios for remote sensing image interpretation, two fundamental challenges constrain the effectiveness of federated fine-tuning: the limited computational capacity of edge devices restricts the multi-scale feature learning capability of lightweight deployed models, while the highly heterogeneous and imbalanced data distributions (Non-IID settings) across clients render conventional parameter aggregation strategies ineffective.
Key Innovation: Extensive experiments on scene classification, object detection, and semantic segmentation benchmarks demonstrate that Fed-RSAdapter achieves an average improvement of 3.75% in overall accuracy for scene classification, 0.74% in mAP for object detection, and 0.25% in overall accuracy for semantic segmentation over federated baselines, while reducing the volume of transmitted parameters to approximately 4.93% of that required by conventional full-parameter federated learning.
68. Category-Aware Global–Local Semantic Alignment for Remote Sensing Image–Text Retrieval
Core Problem: To address these bottlenecks, this study proposes a Category-aware Global–Local Semantic Alignment (CGLSA) framework fine-tuned on the CLIP (ViT-B/16) backbone.
Key Innovation: Compared to strictly controlled CLIP-family baselines under equivalent supervised conditions, CGLSA achieves new state-of-the-art performance across all R@K metrics and mean recall.
69. Target Detection for UAV Inspection in Complex Illumination Environments: A Lightweight Dual-Modal Fusion Method with Multi-Directional Mamba Cross-Attention
Core Problem: Unmanned aerial vehicles (UAVs) frequently operate under complex illumination conditions during urban and industrial security patrols, including low light, fog, and strong backlighting.
Key Innovation: On the OPVM-VIRD dataset, it achieves an mAP50 of 95.7%, which is 2.2% higher than the best result obtained with other fusion methods.
70. Crop-Tool-Augmented Active Perception with Reinforcement Learning for High-Resolution Remote Sensing Visual Question Answering
Core Problem: However, existing vision–language models usually rely on fixed global image inputs, which may lose critical local details in ultra-high-resolution imagery and struggle with sparse informative regions, large object-scale variations, and complex spatial layouts.
Key Innovation: Experiments on VRSBench, MME-RealWorld-RS, XLRS-Bench, and LRS-VQA demonstrate that the proposed method achieves competitive overall performance compared with closed-source, open-source, and remote-sensing-specific vision–language models, and obtains the best or comparable results on most benchmarks.
71. Consistent-Innovation-Aided Distributed Cooperative Localization for Multi-UAV Navigation in GNSS-Denied Environments
Core Problem: However, centralized cooperative localization and exact cross-covariance-based distributed methods usually require global state management or explicit propagation of inter-node cross-covariance, resulting in heavy communication and computational burdens.
Key Innovation: The results show that the proposed method reduces the position, velocity, and yaw RMSEs by 27.14%, 21.90%, and 9.68%, respectively, compared with the non-cooperative method.
72. Water Anomaly Type Identification Based on Deep Multisphere Decision Boundaries Using Remote Sensing Data
Core Problem: Rapid and accurate identification of these events is critical for early warning and informed environmental decision-making.
Key Innovation: Therefore, in this study, a deep multisphere decision boundaries (DMSDB) method for unified, fine-grained identification of diverse water anomalies is developed.
73. Unraveling volcanic cooling in semi-arid climates: model-observation gaps and multi-eruption impacts over Iran
Core Problem: Explosive volcanic eruptions are among the strongest natural climate drivers, yet their regional impacts remain poorly constrained in semi-arid regions.
Key Innovation: The authors present the first comprehensive assessment of volcanically induced cooling over Iran using station observations and CMIP6 simulations for the eruptions of Agung (1963), Fuego (1974), El Chichón (1982), and Pinatubo (1991).
74. A deformation prediction model of high-grade embankments on permafrost considering the pot cover effect
Core Problem: The study quantifies for the first time the significant and escalating role of the PCE in frost heave progression, highlighting the need for combined thermal and vapor-control measures in permafrost embankment design.
Key Innovation: The results demonstrate that the established THVM coupled model effectively characterizes the migration processes of both liquid water and water vapor, as well as the associated consolidation deformation.
75. Loess-like deposits formation dynamics in ephemeral stream valleys of Western Transbaikalia during Late Glacial and Holocene
Core Problem: The relative roles of permafrost degradation, extreme floods and land use in shaping ephemeral valleys through the Late Glacial and Holocene remain difficult to separate.
Key Innovation: Sedimentology and radiocarbon dating resolve three incision phases and show a recent shift toward gullying and accelerated erosion, with human disturbance becoming dominant only during the past 250 years.
76. Modeling-based assessment of channel stability in an anabranching reach of a lowland meandering river
Core Problem: Results indicate that over time, all channels in the anabranching reach evolve toward stability; however, the meandering reach downstream of the anabranching reach exhibits an increase in net deposition over time.
Key Innovation: The dominant mechanism of anabranch development is not documented directly, but appears to be through enlargement and incision of floodplain secondary channels.
77. An improved 15-year record of ice sheet elevation from CryoSat-2 radar altimetry
Core Problem: The measurement accuracy depends critically on the Level-2 processing chain applied to convert the detected surface response to an elevation estimate.
Key Innovation: Here, we present and assess a new and improved 15-year record of ice sheet elevation and elevation change, derived from CryoSat-2 measurements acquired between 2010 and 2025.
78. UniOrtho: From view-centric to unified orthographic framework for satellite semantic 3D reconstruction
Core Problem: Moreover, they achieve only weak coupling between height and semantics via shared encoders or unidirectional feature flow, and lack explicit constraints to enforce distributional consistency.
Key Innovation: Digital Surface Models (DSMs) and Land Cover Maps (LCMs) are fundamental orthographic products for Earth observation applications.
79. Multimodal flow matching for large-scale human mobility flow generation using satellite imagery and social sensing data
Core Problem: However, the acquisition of high-quality human mobility data remains challenging due to high collection costs, limited accessibility, and sparse spatial coverage, motivating the adoption of generative AI to produce representative and scalable human mobility data.
Key Innovation: In this research, we propose a generative GeoAI-based multimodal flow matching method for large-scale origin-destination (OD) mobility flow generation by fusing remote sensing and social sensing data.
80. Theoretical Methods and Practical Applications of Control-Free High-Precision Positioning for Satellite Imagery Aided by Optical Axis Measurement Data
Core Problem: However, affected by the number of ground calibration fields and weather conditions such as clouds and fog during imaging, it is generally difficult to acquire high-quality calibration field images at high frequency, and thus it is hard to perform on-orbit geometric calibration frequently.
Key Innovation: To address this issue, a full-optical-path real-time monitoring method for camera optical axis drift was developed and applied to the engineering practice of the Gaofen-14 (GF-14) satellite.
81. On-orbit relative radiometric calibration of remote sensing satellite planar array sensors with geometric constraints and multi-scale radiometric cascades
Core Problem: Planar array sensors in remote sensing satellites now employ millions of imaging detectors - a hundredfold increase over linear push-broom sensors - creating unprecedented calibration challenges.
Key Innovation: The study develops an efficient, high-accuracy on-orbit relative radiometric calibration method for planar array sensors that overcomes these limitations.
82. Multitemporal latent dynamical framework for hyperspectral images unmixing
Core Problem: However, this motivation is hindered by two challenges: the inherent complexity in defining, modeling and solving problem, and the absence of theoretical support.
Key Innovation: To address above challenges, in this paper, we propose a multitemporal latent dynamical (MiLD) unmixing framework by capturing dynamical evolution of materials with theoretical validation.
83. Label-efficient mapping of unregulated waste dumps via mixed supervision with feature masked recovery
Core Problem: Mapping unregulated solid waste dumps from remote sensing imagery is important for environmental risk assessment and sustainable development monitoring, but remains constrained by the high cost of pixel-level annotation.
Key Innovation: Experiments on the newly constructed YREB dataset and the enhanced Global Dumpsite benchmark show that the proposed method improves over few-shot, pure weakly supervised, and semi-supervised label-efficient baselines under spatially independent validation, cross-dataset transfer, and conventional patch-level benchmarks.
84. Referring Remote Sensing Image Segmentation method based on Scene-Aware Guided Network model
Core Problem: However, existing RRSIS approaches are limited to pixel-level cognitive abilities, failing to effectively handle the diverse geographical features inherent in remote sensing imagery, which significantly impairs their segmentation performance across different geographical scenes.
Key Innovation: Experimental evaluation on RefSegRS, RRSIS-D, and LandRef datasets reveals that the proposed SAGNet method consistently outperforms existing RRSIS approaches.
85. A novel edge-based accuracy assessment framework for target detection from remotely sensed images
Core Problem: Accuracy assessment is critical for evaluating map quality and improving algorithms for target detection in remotely sensed images.
Key Innovation: The proposed framework therefore provides a new capability for assessing the accuracy of target detection maps, which is of vital importance for evaluating and improving state-of-the-art remote sensing products.
86. An optimization AI model based on MoE for retrieving land surface temperature and emissivity
Core Problem: The main drawbacks of contemporary MODIS land surface temperature (LST) and emissivity (LSE) outputs stem from cloud pollution, imprecision in atmospheric moisture vapor rectification, preliminary miscalculations of variable surface emissivity, and heterogeneity at the sub-pixel level, leading to significant potential for advancements in the reliability of these outputs.
Key Innovation: The study proposes an iterative optimization strategy based on an Optimization Artificial Intelligence Model Based on Mixture of Experts (Opt-MoE) model.
87. Robust hyperspectral anomaly detection via low-rank discriminative dictionary and global-local saliency weight
Core Problem: However, existing methods are limited in detection performance due to single constraints, insufficiently robust dictionaries, and inadequate feature utilization.
Key Innovation: Experimental results on this dataset and three publicly datasets demonstrate that the proposed method outperforms other classical and advanced algorithms, exhibiting its robustness and adaptability.
88. PMTB-VSRnet: A video super-resolution network with radiometric consistency guidance for Antarctic passive microwave brightness temperature
Core Problem: Monitoring Antarctic surface melt using satellite-based Earth observation relies heavily on passive microwave data for their all-weather capability; however, their coarse spatial resolution limits the detection of fine-scale melt features.
Key Innovation: To address this limitation, we propose PMTB-VSRnet, a video super-resolution network with radiometric consistency guidance for Antarctic passive microwave data.
89. Reconstructing 500-m NDVI time series since 1984 using a target self-attention spatiotemporal fusion model
Core Problem: A major gap in long-term, large-scale land surface monitoring is the absence of consistent Moderate Resolution Imaging Spectroradiometer (MODIS)-like data prior to its 2000 launch.
Key Innovation: The authors propose a novel Target self-Attention Spatiotemporal Fusion (TASTF) model that bridges the scale gap between Landsat (30 m) and Global Inventory Modeling and Mapping Studies (GIMMS) (∼8 km) data to generate seamless, 500-m biweekly MODIS-like Normalized Difference Vegetation Index (NDVI) data dating back to 1984 over the contiguous United States (CONUS).
90. Mapping slash-and-burn in humid tropical rainforests based on Sentinel-1 imagery and 3D deep learning
Core Problem: However, optical data-based monitoring is limited by insufficient continuous spatio-temporal observations.
Key Innovation: To address this issue, we propose an approach that employs a three-dimensional deep learning architecture to simultaneously extract multi-level embedded spatial context and temporal change characteristics-such as magnitude and duration features from time series S1 data to differentiate single and double disturbances.
91. Overcoming spectral and tidal challenges in mangrove mapping with a multi-modal deep learning network fusing phenological and texture features
Core Problem: Accurate mangrove mapping faces persistent challenges from spectral confusion and tidal fluctuations.
Key Innovation: Evaluated across three Chinese coastal sites, our method achieved superior overall accuracies (95.14%, 94.21%, and 96.66%), outperforming benchmarks by 0.78-8.60%.
92. Microtopographic and runoff-sediment responses to check-dam configurations under consecutive rainfall: Laboratory experiments in a simulated loess watershed
Core Problem: Check dams are widely used in the loess hilly region to reduce soil erosion and sediment export, yet how different dam configurations influence runoff-sediment processes and microtopographic adjustment during consecutive storms remains poorly understood.
Key Innovation: Overall, the results indicate that dam placement affects both sediment redistribution and microtopographic development during consecutive rainfall.
93. Assessment and prediction of snow cover dynamics over Siachen using advanced statistics and Hybrid Seasonal-Trend Fourier Regression model
Core Problem: The Siachen Glacier, one of the largest high-altitude glaciers in the Himalayas, plays a critical role in regional hydrology and climate regulation.
Key Innovation: The study employs Sentinel-2 Level-2 A imagery from 2022 to 2025 and the Normalized Difference Snow Index (NDSI) to generate monthly snow cover maps and assess spatiotemporal variability across the glacier.
94. Automatic identification and filtration of individual tectonic discontinuities in rock mass using improved DBSCAN
Core Problem: Nonetheless, the accurate identification of individual tectonic discontinuities remains challenging due to the difficulty in distinguishing between blast-induced cracks and original tectonic discontinuities on rock tunnel faces.
Key Innovation: The study proposes a new automatic algorithm that uses flatness evaluation to obtain core points for filtering the point cloud, which is subsequently employed for clustering.
95. Urban geological assessment for SDG 11 alignment: empirical evidence from 11 cities in Zhejiang, China
Core Problem: Sustainable Development Goal 11 (SDG 11) emphasizes building inclusive, safe, and resilient cities; however, the integration of urban geological systems into risk-informed governance to advance this goal remains understudied.
Key Innovation: Focusing on Zhejiang Province, a rapidly urbanizing coastal region characterized by complex geological constraints, this study develops a novel multidimensional evaluation framework to assess urban geological resilience across four dimensions: Environment, Resources, Safety, and Data Digitalization, and systematically mapping each indicator to specific SDG 11 sub-targets.
96. Intelligent rock classification from a mesoscopic perspective: A computer vision approach based on thin section images
Core Problem: To address the challenges of obtaining macro-level parameters and high costs associated with traditional rock classification methods, this study proposes a novel intelligent rock classification method based on rock thin section (RTS) images and computer vision (CV) from a mesoscopic perspective.
Key Innovation: To address the challenge of scarce physical samples, the Lanczos interpolation algorithm was introduced for data augmentation.
97. Synergistic multi-sensor approach for soil moisture estimation in semi-arid monsoonal basins
Core Problem: Soil moisture regulates land-atmosphere coupling, agricultural productivity, hydrological processes, and ecosystem functions; however, accurately estimating it over heterogeneous semi-arid regions remains challenging.
Key Innovation: The present results demonstrate that an interpretable, multi-sensor integration approach can resolve regional-scale soil moisture dynamics in semi-arid environments and support land and water resource management in areas with sparse in situ observations.
98. A four-decade satellite record of coupled lake area and water quality responses to climate changes and human activities in Northwest China
Core Problem: Understanding the coupled dynamics of lake (including natural lakes and reservoirs) water quantity and quality remains scarce, both regionally and globally.
Key Innovation: The study establishes a novel remote sensing-based framework to assess coupled lake quantity-quality dynamics, represented by lake area and Secchi disk depth (SDD), and to classify lakes into four ecological-hydrological trajectories (Types I-IV).
99. Tracing hydrogeological sources of spring discharges to peat-forming bofedales in the Peruvian Andes
Core Problem: Peat-forming bofedales are unique environments found at high altitudes across the Peruvian Andes that provide critical ecosystem services such as sustaining streamflow to downslope communities.
Key Innovation: As limited studies have attempted to address seasonally dynamic groundwater flow from springs, we conducted a hydrogeophysical investigation to characterize and quantify specific discharges from springs over two weeks following the ‘wet-up’ period of two hydrogeologically distinct bofedales in the Upper Ramuschaka Watershed (Cusco, Perú).
100. Numerical Simulation of Cone Penetration Test in Tailings Sand Using a State-dependent 3D Elastoplastic Model
Core Problem: In this study, a state-dependent three-dimensional elastoplastic constitutive model of tailings sand was developed based on critical state theory and implemented into ABAQUS through a user-defined material subroutine.
Key Innovation: The results indicate that the cone penetration resistance is strongly governed by the state-dependent mechanical behavior of tailings sand and the development of the surrounding plastic zone.
101. Neural-BGC: An Observation-Driven Emulator for Hybrid Physical-Biogeochemical Modeling
Core Problem: Coupled ocean‐biogeochemical models are essential for understanding marine ecosystems, yet they often suffer from persistent biases due to poorly constrained empirical parameters.
Key Innovation: To address this limitation, we introduce Neural‐BGC, an observation‐driven neural network emulator for hybrid physical‐biogeochemical modeling.
102. HPC-Enabled Video-based Coastal Wave Parameter Estimation Using V-JEPA and Deep Spatiotemporal Learning
Core Problem: High deployment cost, poor spatial coverage and susceptibility to storm conditions are all challenges faced by traditional in-situ methods.
Key Innovation: The study presents a video-based and high performance computing (HPC) enabled deep learning framework for joint sensor free estimation of five coastal wave parameters, namely significant wave height (Hs), maximum wave height (Hmax), peak period (Tp), zero upcrossing period (Tz) and wave direction (theta) from monocular coastal video.
103. A hybrid analytical-PINN model for subsurface simulation of geothermal heat exchangers in heterogeneous underground
Core Problem: In this paper, a parametric physics-informed neural network for solving the heterogeneous soil thermal problem with borehole heat exchangers (BHEs) as singular sources is developed.
Key Innovation: There are three novel features in the present framework; namely, (i) the singularity is naturally removed by using analytical line source models; (ii) using the explicit formulation for gradient thermal conductivity enables physics-informed learning of the parametrization featuring the conductivity; (iii) the learned correction is utilized as an efficient universal corrector via superposition principles.
104. GeoSEAN: Explainable Country-Level Image Geolocation for ASEAN Regions
Core Problem: However, this task remains challenging in regions where countries share similar urban, roadside, architectural, and environmental characteristics.
Key Innovation: These results demonstrate that the proposed model can support accurate regional image geolocation while enabling object level inspection of the visual cues underlying its predictions.
105. UMSS: Towards Unsupervised Multi-modal Semantic Segmentation
Core Problem: Multimodal semantic segmentation (MSS) is essential for robust perception in complex environments, yet its potential remains largely untapped because of the prohibitive cost of human annotations.
Key Innovation: To this end, we propose UniM2 (Unified Multimodal), a novel framework built on DINOv3 that transforms conventional fusion methods into consistent performance gains.
106. Exploring Zero-Shot Foundation Models for Multivariate Time Series Anomaly Detection
Core Problem: However, these observations yield valuable insights: the error peaks at anomaly boundaries, indicating FMs reliably detect distribution changes.
Key Innovation: The authors investigate the zero-shot application of a univariate forecasting FM, TimesFM, to industrial MTSAD on the Secure Water Treatment (SWaT) benchmark, evaluating two strategies: treating the FM as a per-feature forecaster with thresholded prediction errors, and as an embedder whose intermediate representations feed standard outlier detectors.
107. Weakly Supervised Spatio-Temporal Candidate Discovery of Dairy Farm Sites from Seasonal Satellite Imagery
Core Problem: Incomplete labels make it difficult to discover spatial targets whose evidence is distributed across structures, land cover and seasonal change.
Key Innovation: A weakly supervised pipeline combines self-supervised seasonal Sentinel embeddings, open-map priors and graph smoothing; among 26,722 tiles, it ranks 71 candidate clusters and reaches 0.80 precision within 1 km for the top five.
108. RFMSR: Residual Flow Matching for Image Super-Resolution
Core Problem: Recent flow-matching-based vision-only approaches have made significant strides; however, they adopt standard flow formulations that transport from a pure Gaussian prior to the data distribution, discarding the rich structural information already present in the low-quality (LQ) input.
Key Innovation: The authors further introduce a two-phase training strategy: Phase I pretrains the velocity field via conditional flow matching, while Phase II applies end-to-end supervision to the single-step prediction while retaining the velocity loss across all timesteps, achieving high-quality single-step generation without sacrificing multi-step refinement.
109. Ensemble Controlled-Flow Filtering for Implicit Data Assimilation
Core Problem: Many observation mechanisms, however, are many-to-one, implicit, non-smooth, or accessible only through simulation, and need not provide the residual structures or likelihood guidance required by existing ensemble filters.
Key Innovation: Numerical results show that Kalman-type filters remain preferable for smooth additive-Gaussian observations, while the proposed methods are better suited to non-Gaussian, many-to-one, multimodal, and implicit observation models.
110. Sat2RealCity: Geometry-Aware and Appearance-Controllable 3D Urban Generation from Satellite Imagery
Core Problem: Supported by our constructed BuildVerse3D dataset, (1) we introduce an OpenStreetMap (OSM)-guided spatial grounding strategy to inject geospatial constraints into the 3D generation process; (2) we design an appearance-guided controllable generation mechanism for realistic architectural appearance and regional style consistency; and (3) we construct an MLLM-powered semantic pipeline for regional appearance understanding and semantic-aware appearance synthesis.
Key Innovation: To address these limitations, we present Sat2RealCity, a grounded urban generation framework that leverages object-level 3D generative priors for scalable city synthesis from satellite imagery.
111. GAINS: Gaussian-based Inverse Rendering from Sparse Multi-View Captures
Core Problem: However, the accuracy of these methods degrades under sparse-view settings, where limited observations lead to severe ambiguity between geometry, reflectance, and lighting.
Key Innovation: The authors introduce GAINS (Gaussian-based Inverse rendering from Sparse multi-view captures), a two-stage inverse rendering framework that leverages foundation models as priors to stabilize geometry and material estimation.
112. MUSA-PINN: Multi-scale Weak-form Physics-Informed Neural Networks for Fluid Flow in Complex Geometries
Core Problem: The locality bias of point-wise constraints fails to propagate global information through tortuous channels, causing unstable gradients and conservation violations.
Key Innovation: To address this, we propose the Multi-scale Weak-form PINN (MUSA-PINN), which reformulates Navier-Stokes equation constraints as integral conservation laws over hierarchical spherical control volumes.
113. Pixel-Level Pavement Distress Assessment Using Instance Segmentation
Core Problem: The results show that instance segmentation is a practical direction for field pavement imagery and aggregate crack-area estimation, while also exposing open challenges in annotation consistency, class imbalance, confounder rejection, and mask-level benchmarking.
Key Innovation: Automated pavement distress assessment requires more than image-level classification or coarse bounding box detection, demanding precise localization of thin, branching, and irregular cracks to achieve the geometric precision necessary for maintenance-relevant quantification.
114. TRIG: Trajectory-Rig Decoupled Metric Geometry Learning
Core Problem: The authors introduce decoupled pose encoding and supervision, which separately constrain trajectory evolution and rig geometry for metric-consistent learning.
Key Innovation: The authors propose Trajectory-Rig Decoupled Metric Geometry Learning (TRIG), a geometry perception framework for autonomous driving.
115. Numerical simulation of complete stress-strain curves of brittle rocks using 3D FDEM
Core Problem: Parameter calibration in 3D finite-discrete element method (FDEM) simulations remains challenging due to complex micro-macro relationships and the limited consideration of post-peak behavior in existing approaches.
Key Innovation: The results demonstrate that the proposed method provides an efficient and practical approach for parameter calibration in 3D FDEM simulations of brittle rock failure.
116. A skill-enhanced retrieval-augmented generation expert system framework for offshore accident analysis: Combining fuzzy inference, Dempster-Shafer evidential reasoning, and expert-gated skill adaptation
Core Problem: Traditional rule-based expert systems for offshore safety assessment suffer from labour-intensive knowledge acquisition and an inability to evolve with operational experience.
Key Innovation: The framework is positioned as a feasibility-oriented knowledge-engineering contribution: it demonstrates that Skill-structured RAG can support maintainable, modular, and traceable offshore safety knowledge bases, not that it outperforms established expert-assessment methods on general offshore risk problems.
117. A novel approach for efficient ship hydrodynamic simulations with realistic coastal bathymetries
Core Problem: However, the high computational cost of existing high-fidelity numerical models limits their practical use, particularly for simulations involving large domains, complex bathymetries, and long-duration realistic sea conditions.
Key Innovation: Accurate numerical prediction of ship motions and hydrodynamic loads is crucial for the design of energy-efficient and reliable vessels.
118. Statistical analysis and probabilistic modelling of irregular breaking wave forces on a vertical pile within pile groups
Core Problem: This investigation was motivated by a lack of statistical and probabilistic insights for such forces, which represent a limitation for coastal and offshore engineering applications.
Key Innovation: These findings enhance the understanding of group interaction effects and aim to support improved probabilistic design of pile group structures.
119. Revisiting and enhancing methods of representative wave height selection for medium-term coastal bed level evolution applications
Core Problem: Accurate prediction of medium-term coastal bed level evolution via numerical modelling is essential for effective coastal management but remains computationally demanding when process-based models are forced with full wave climate datasets.
Key Innovation: The study systematically revisits classical RMWH approaches and proposes three enhanced variants aimed at improving accuracy without increasing computational cost.
120. DCFNet: Dual-Domain Collaborative Focusing Network for Infrared Small Target Detection
Core Problem: To address the challenges of low target contrast, large scale variations, and insufficient spatial-frequency information utilization in complex backgrounds for infrared small target detection (IRSTD), this article proposes a novel dual-domain collaborative focusing network (DCFNet).
Key Innovation: First, a frequency interaction enhancement module is designed to dynamically balance low-frequency background and high-frequency target components, enabling precise matching between deep semantic features and shallow details, thereby improving the consistency of multiscale feature fusion.
121. Saliency-Modulated Group Dynamic Attention for Hyperspectral Image Classification
Core Problem: Hyperspectral image classification (HSIC) remains challenging due to severe spectral redundancy, complex spatial-spectral dependencies, and the scarcity of labeled samples.
Key Innovation: To address these challenges, this article proposes SGDAFormer, a transformer-based framework for HSIC that incorporates the saliency-modulated group dynamic attention (SGDA) mechanism as its core contribution.
122. WCS-Net: A Wavelet-Guided Spatial-Frequency Fusion Network for Detection-Oriented Sea Clutter Suppression in Marine Radar Remote Sensing
Core Problem: Existing deep restoration networks mainly operate in the spatial domain and often lack explicit modeling of frequency-dependent clutter structures and detection-oriented target preservation.
Key Innovation: Experiments on real measured marine radar data with simulated targets show that the default WCS-Net achieves the best overall detection-oriented performance among representative baseline methods on the evaluated datasets.
123. Mapping Urban Socioeconomic Disparities Using Remote Sensing Hyperspectral Data
Core Problem: Socioeconomic information is foundational for sustainable urban development; however, conventional household surveys remain labor-intensive, infrequent, and spatially inconsistent-particularly in rapidly growing cities in India.
Key Innovation: With the increasing availability of hyperspectral data, this study aims to address a key question: Can it be utilized to generate socioeconomic maps with improved accuracy?
124. Prior Density Map Guided Feature Modulation for Ship Detection in SAR Imagery
Core Problem: Ship detection in synthetic aperture radar (SAR) imagery is a critical task for maritime surveillance, but it still faces challenges such as small-sized targets, strong nearshore clutter, and limited discriminative information from single-modal SAR data.
Key Innovation: Most existing deep learning methods improve detection by modifying network structures or introducing attention mechanisms, while ignoring valuable spatial distribution priors of ship occurrences.
125. Impact of Interseasonal Climate Variation on Crop Classification With Deep Learning and Sentinel-1 Data in Denmark
Core Problem: However, increasing interseasonal climate variation alters crop phenology, raising concerns about the generalization of models trained on past data under rapidly changing climatic conditions.
Key Innovation: Deep learning methods, when combined with satellite remote sensing data, have emerged as powerful tools to deliver unprecedented accuracy and scalability for crop mapping.
126. Complex Scattering-Aware and Globally Enhanced Hybrid Network for SAR Target Recognition
Core Problem: However, existing complex-valued neural networks largely neglect the global structure of SAR targets and the adaptive modeling of physical scattering mechanisms across different scales.
Key Innovation: Experimental results on the MSTAR and OpenSARShip datasets demonstrate that the proposed CSGEHNet achieves notable improvements in SAR target recognition accuracy.
127. Hyperspectral Image Classification Using a Hybrid Vision Transformer and Graph Convolutional Network
Core Problem: Hyperspectral image classification (HSIC) plays a critical role in agriculture, environmental monitoring, and urban planning.
Key Innovation: The proposed architecture incorporates: 1) a self-attention-driven dynamic graph for capturing fine-grained interpixel relationships, 2) adaptive multiscale patch embedding for improved scene variability handling, and 3) streamlined Transformer-GCN modules optimized through pruning and distillation for computational efficiency.
128. Fit-for-purpose assessment of satellite aerosol and cloud datasets for constraining and monitor aerosol–cloud interactions
Core Problem: Aerosol-cloud interactions (ACI) remain one of the dominant sources of uncertainty in estimates of anthropogenic effective radiative forcing.
Key Innovation: Within the framework of the "Satellite observations to improve our understanding of aerosol-cloud interactions" (SATACI) project, we perform a comprehensive fit-for-purpose (F4P) assessment of existing satellite aerosol and cloud datasets used to (i) quantify aerosol indirect effects on liquid and mixed-phase clouds and (ii) support the feasibility study of a novel aerosol-cloud climate indicator.
129. Reduced tropical rain-band asymmetry through the seasonal cycle
Core Problem: Enhanced austral-summer heating raises southern tropical sea surface temperatures above a convective threshold, triggering a transient southern rain band amplified by wind-evaporation-surface temperature feedback, while convection remains suppressed outside this season south of the Equator.
Key Innovation: The rain-belt position thus reflects the nonlinear integration of seasonal dynamics rather than a simple response to annual-mean forcing.
130. Spectral-Prior-Guided Swin TransUnet for Sparse-Aperture FMCW MIMO-SAR Imaging
Core Problem: In millimeter-wave frequency-modulated continuous-wave (FMCW) multiple-input multiple-output synthetic-aperture radar (MIMO-SAR) imaging, platform displacement beyond the spatial Nyquist limit during a slow-time sampling interval creates aperture gaps, causing azimuth aliasing and degraded resolution.
Key Innovation: The study proposes a spectral-prior-guided Swin TransUnet (SSTU) method for suppressing azimuth ambiguity in sparse moving-array imaging.
131. A Method for Extracting Small Moving Targets Based on Temporal and Spatial Saliency Features
Core Problem: To address the problem of detecting dim small targets against complex moving backgrounds, based on spatiotemporal saliency features, this paper proposes a novel extraction model: the temporal and spatial saliency feature (TSSF) model.
Key Innovation: In the temporal domain processing stage, a temporal motion saliency operator is constructed by leveraging the differences in temporal motion responses between the target and the background; this operator simultaneously filters for candidate target points while suppressing large-area background clutter.
132. IC-EWH: Energy-Weighted Hough Transform with Iterative Curvature Compensation for Squint Angle Estimation of Highly Squinted SAR
Core Problem: However, range curvature hinders its estimation accuracy.
Key Innovation: To solve the above problem, this paper proposes a novel squint angle estimation scheme.
133. HFG-YOLO: A High-Frequency-Guided Network for Small-Object Detection
Core Problem: Small objects in UAV and remote sensing images often occupy only a few to several dozen pixels, and their edge and texture information is easily further weakened after processing by the network backbone.
Key Innovation: On VisDrone2019, HFG-YOLO achieves a 31.10% mAP50 and 6.08% mAPs, outperforming the YOLO11n baseline by 2.70 and 1.71 percentage points, respectively.
134. Significance-Preserving Progressive Network for Infrared and Visible Image Fusion
Core Problem: However, existing methods often struggle to balance global dependency modeling with local detail preservation and to effectively coordinate heterogeneous local and global features during fusion.
Key Innovation: Extensive experiments on the MSRS, RoadScene, and TNO datasets demonstrate that SiPFusion achieves competitive visual quality and strong overall quantitative performance against 15 state-of-the-art fusion methods, obtaining leading results on most evaluated metrics.
135. CASA-Net: Context-Aware Small-Object Adaptation Network for UAV Aerial Images
Core Problem: Detecting small targets in UAV aerial imagery is inherently difficult because these objects occupy only a small number of pixels and are highly susceptible to cluttered backgrounds, dense spatial arrangements, and pronounced scale variation.
Key Innovation: To address this problem, we propose CASA-Net (Context-Aware Small-object Adaptation Network), a context-aware detector built on a YOLOv26s baseline with three coordinated improvements: an Enhanced Small-Target-Aware Label Assignment mechanism for stronger supervision of tiny instances, a Multi-scale Feature Enhancement Module for richer contextual representation and spatial discrimination, and an aerial-specific augmentation pipeline for improved robustness to viewpoint, scale, and motion blur.
136. Influence of Internal Climate Variability on Satellite-Altimeter-Derived Regional Sea-Level Trends
Core Problem: Regional sea-level trends derived from satellite altimetry deviate substantially from the global mean, but the relative roles of externally forced change and internally generated climate variability remain difficult to separate from the short satellite record.
Key Innovation: It does not include DUACS mapping errors, inter-mission calibration uncertainty, geophysical correction uncertainty, glacial-isostatic-adjustment-related bias, or uncertainty in the forced sea-level response.
137. Multi-Stage Aggregation CNN–Transformer Hybrid Architecture for Infrared Small-Target Detection
Core Problem: Infrared targets with weak contrast and few pixels are easily lost when a detector cannot preserve fine detail while modelling broader scene context.
Key Innovation: MACT-Net combines multilevel aggregation, CNN-Transformer feature modelling and vector-quantized bottleneck compression, with tests on three public datasets isolating the contribution of each component.
138. Short-Term Forecasting of Ocean Surface Current Maps Using High-Frequency Radar Observations and LSTM Neural Networks: A Case Study of Southwestern Taiwan
Core Problem: Short-term coastal surface-current forecasts at forecast lead times (τ = 1–12 h) are critical for search and rescue (SAR), pollution response, and vessel routing.
Key Innovation: From a 2015–2019 archive of hourly CODAR high-frequency radar (HFR) observations off Southwestern Taiwan, we developed grid-point long short-term memory (LSTM) models using historical observations alone, without atmospheric forcing or data assimilation (training 2015–2017, validation 2018, test 2019).
139. VODet: A Vertex Offset-Based Method for Oriented Object Detection in Remote Sensing Images
Core Problem: To tackle the challenges in oriented object detection, such as discontinuous angle regression, low precision in discrete classification, and high complexity in probabilistic modeling, this paper proposes VODet (Vertex Offset-based Detector), a novel oriented object detection method.
Key Innovation: Extensive experiments on DOTA, HRSC2016, and UCAS-AOD demonstrate that VODet achieves competitive detection accuracy (80.44% mAP on DOTA, 90.42% AP on HRSC2016, and 90.11% mAP on UCAS-AOD), with particular advantages for objects of large aspect ratios and dense arrangements.
140. A Novel Target Extraction and Energy-Balancing Method for HoloSAR 3D Imaging
Core Problem: However, after conventional subaperture-wise TomoSAR reconstruction and non-coherent integration, the resulting 3D imagery suffers from severe dynamic range imbalance due to angle-dependent scattering responses: wide-angle strong scatterers are repeatedly amplified, whereas narrow-angle weak structures are buried below the noise floor.
Key Innovation: Quantitative evaluation on Ku-band UAV circular SAR data demonstrates that the proposed method improves the Target-to-Background Ratio by 0.7 dB (to 11.2 dB), achieves a Background Suppression Ratio of −5.2 dB, increases the Structural Completeness Index by 156% (to 1428.1), and compresses the original dynamic range imbalance, which exceeds 50 dB, while preserving scene physical realism (ENL ≈ 7.4 × 10−3).
141. A Scale-Invariance-Based Algorithm Application for Land Surface Temperature Downscaling in Denmark
Core Problem: Note that while several Machine Learning (ML) models besides Linear Regression (LR) were considered for the MTS architecture, only LR was used for the STS one due to the limited amount of available data which the former require for hyperparameter tuning.
Key Innovation: Herein, two scale-invariance-based architectures were developed: a single-timestamp (STS) model, trained with coarse data of the timestamp whose fine target it infers; and a multi-timestamp (MTS) one, trained with multiple timestamps.
142. Cross-Platform Comparison of Marine Boundary Layer Cloud and Drizzle Properties over the Southern Ocean Using Airborne, Shipborne, and Satellite Observations
Core Problem: Marine boundary layer (MBL) clouds strongly influence radiation and precipitation over the Southern Ocean (SO), yet their vertical structures and microphysical properties remain poorly constrained across observational platforms.
Key Innovation: An empirical reflectivity–microphysics retrieval framework developed from in situ droplet size distributions (DSDs) measured during SOCRATES was applied to MARCUS M-WACR and CloudSat CPR reflectivity observations to retrieve vertical profiles of number concentration (N), effective radius (re), and liquid water content (LWC) for cloud and drizzle particles.
143. Spatial mapping of environmental risk in riverine wetlands: a case study of Crete, Greece
Core Problem: Wetlands, including associated waterbody types such as riverine wetlands, are highly vulnerable to pollution and ecological degradation driven by surrounding land-use activities.
Key Innovation: The study presents a risk-mapping methodology designed to predict where such impacts are most likely to occur.
144. Scalable physics-informed learning for Allen-Cahn equations
Core Problem: Physics-informed learning has been increasingly explored in geotechnical engineering, but its computational efficiency and practicability in engineering practice often incur scepticism.
Key Innovation: To address these limitations, this study enhances current physics-informed learning by integrating domain decomposition, transfer learning, data assimilation and customised optimiser and loss functions.
145. Thermo-Mechanical Effects on the Mechanical Anisotropy and Fracture Evolution of Bedded Sandstones
Core Problem: The results show that thermally induced microcracks redistribute from the bedding zones to the rock matrix as temperature increases, indicating a critical temperature threshold.
Key Innovation: In this study, a numerical model of bedded sandstone was developed using the Particle Flow Code in two dimensions (PFC 2D ), incorporating temperature, confining pressure, and bedding angle.
146. Mechanical Behaviour of Gap-graded Soils Subjected to Particle Removal: Influence of Fines Content and Particle Size Ratio
Core Problem: Particle loss can reorganize gap-graded soils, but its effects on strength, volume change and drainage response depend jointly on fines and particle-size ratio.
Key Innovation: Ninety drained and undrained triaxial tests use salt dissolution to control particle removal; higher fines and lower size ratios increase deformation and reduce strength, while neural sensitivity analysis identifies fines content and void ratio as dominant controls.
147. Structural impacts under shared regional forcing in sandy beach systems
Core Problem: However, beaches exposed to comparable regional forcing can exhibit contrasting morphodynamic responses.
Key Innovation: Alongshore coherence introduces a third, partially independent dimension of response, distinguishing between integrated and fragmented behaviour that is weakly related to forcing magnitude and strongly site-dependent.
148. Bedrock river channel width in the Western Qinling range: Spatial variation characteristics, controls and implications
Core Problem: Bedrock-channel width responses to uplift, lithology and transient erosion are less well constrained than channel-slope adjustment.
Key Innovation: Field measurements and terrain analysis across nine Western Qinling catchments show that relative uplift outweighs lithology in several reaches and that channel narrowing begins upstream of knickpoints before slope adjustment.
149. A systemic framework for the assessment of water resilience in healthcare facilities
Core Problem: Healthcare facilities constitute critical infrastructures whose operational continuity is closely dependent on the availability of potable water, a key enabling resource for essential clinical, diagnostic and logistical processes.
Key Innovation: The study develops a methodological decision-support framework for assessing hospital water resilience as an interdependent socio-technical system.
150. Assessing Individual Level Power Outage Vulnerability through Synthetic Populations
Core Problem: The study addresses this gap by developing a comprehensive power outage vulnerability dataset for Houston, Texas, integrating synthetic population modeling with census tract-level outage likelihood focusing on Winter Storm Uri.
Key Innovation: Using Public Use Microdata from the American Community Survey (2015-2019) and Oak Ridge National Laboratory’s UrbanPop framework, we constructed a synthetic population to evaluate intersecting social vulnerabilities at the individual level.
151. Uncertainty reduction in quantitative characterization of ablating river ice microstructure for reliable ice-load assessment
Core Problem: The structural reliability of hydraulic engineering and offshore structures in cold regions depends critically on the accurate assessment of ice loads.
Key Innovation: Validation against physical density measurements demonstrates that the proposed method significantly outperforms traditional approaches, reducing the mean deviation to merely 1%.
152. Vulnerability analyses of power-water interdependent infrastructure networks by using load-functional dependency coupled cascading failure model
Core Problem: Previous studies on vulnerability analyses of interdependent critical infrastructure networks are mainly based on cascading failure models that only consider the effect of network topologies and binary dependencies, overlooking progressive functional degradation.
Key Innovation: The study proposes a novel cascading failure model that explicitly couples load dynamics with progressive functional dependency in power-water interdependent infrastructure networks.
153. Seasonal mass balance of Hofsjökull ice cap, Iceland, estimated from repeat ICESat-2 satellite altimetry
Core Problem: Glaciological in-situ measurements have been conducted for almost four decades to monitor the mass balance of Icelandic glaciers, but they are potentially biased due to the limited spatial coverage of the stake network.
Key Innovation: The study presents a novel framework for estimating the seasonal mass balance of the Hofsjökull ice cap in Iceland using 91-day repeat elevation measurements from ICESat-2 satellite altimetry and linear-spline hypsometric interpolation.
154. Optimizing 1D-Var near-surface wind speed from SSMIS with RapidScat
Core Problem: An advantage of the presented scatterometer and radiometer collocation methodology is that the need for calibration adjustment can be evaluated, ensuring consistency of wind speed from MWI with SCA onboard EPS-SG.
Key Innovation: The compatibility of passive microwave wind speed retrievals with scatterometer winds is investigated to prepare for the Second Generation EUMETSAT Polar System Second Generation (EPS-SG) satellites.
155. Optimized retrievals of aerosol optical properties from directional polarimetric camera using optimal linear mixture of basis aerosol models supported by the non-negative matrix factorization
Core Problem: This poses a challenge to improving retrieval accuracy and expanding the range of retrievable parameters.
Key Innovation: To address this issue, this study proposes an optimal aerosol model construction method based on the non-negative matrix factorization (NMF) approach.
156. Retrieval of ocean particle backscatter coefficient profile based on Spaceborne Active-Passive Fusion Fernald Algorithm
Core Problem: However, current retrieval algorithms are significantly affected by the accuracy of wind speed and the ambiguous relationship between the 180° particle volume scattering coefficient and b bp .
Key Innovation: To address these challenges, the Spaceborne Active-Passive Fusion Fernald Algorithm (SAPFFA) was developed in this study.
157. Road extraction for complex urban interchanges by integrating topological structure and geometric semantics
Core Problem: Interchanges are critical transportation infrastructures for alleviating urban traffic congestion.
Key Innovation: Furthermore, to improve feature discriminability and highlight road structures, a diversified attention-based optimization and enhancement module is introduced.
158. SGMS-Fusion: Semantic guided multi-scale fusion towards accurate vehicle localization via ground-aerial cross-view matching
Core Problem: Achieving precise localization of vehicles remains a formidable challenge in urban areas where GPS signals suffer from occlusion and multi-path interference caused by high-rise structures.
Key Innovation: The study presents a new cross-view geo-localization framework which aligns street view images with aerial imagery while incorporating semantic segmentation.
159. Dim moving target detection with hyperspectral image sequences
Core Problem: Dim moving target detection (DMTD) in hyperspectral image sequences (HSIS) aims to identify potential small moving anomalous targets with low contrast relative to the background of HSIS and has garnered substantial attention in various remote sensing photography and surveying applications.
Key Innovation: (2) For effective anomaly suppression during background reconstruction, we design a Haar discrete wavelet transform convolution module that explicitly captures the discriminative frequency characteristics between targets and background, significantly improving detection accuracy.
160. Agricultural parcel polygon reconstruction using boundary line segment assembly
Core Problem: However, these pipelines often result in limited boundary regularity, redundant polygon vertices, and incomplete reconstruction of parcels that span multiple image patches.
Key Innovation: To address these issues, we propose the Boundary-to-Polygon Network (B2PNet), a Boundary Line Segment-based (BLS-based) reconstruction pipeline for generating agricultural field polygons.
161. CAH-VAE: Channel-aware hierarchical VAE for hyperspectral image compression
Core Problem: However, hyperspectral images are compressed into a compact latent variable representation in low-dimensional spaces, the entanglement of spatial and spectral features poses a challenge, making it difficult to preserve fine spatial details and spectral fidelity at low bitrates.
Key Innovation: To address this, we propose a Channel-Aware Hierarchical Variational Autoencoder (CAH-VAE) for high-fidelity satellite hyperspectral image compression at ultra-low bitrates.
162. Motion consistency-guided spatiotemporal learning: a robust framework for low-altitude infrared object detection
Core Problem: However, accurately detecting dim and tiny targets remains highly challenging in complex scenes, where targets are often captured under low signal-to-noise ratio conditions and appear as weak responses embedded in drifting clouds, swaying vegetation, and other background clutter.
Key Innovation: Specifically, we first develop a bio-inspired motion perception mechanism that combines ON/OFF pathway modeling with fractional-order temporal integration to improve sensitivity to weak motion.
163. Geometric positioning and correction for the FY-3F HIRAS
Core Problem: The High-Resolution Infrared Sounding Instrument for Atmospheric Studies (HIRAS) onboard the Fengyun-3 (FY-3) satellite is a key payload for meteorological observation and numerical forecasting in China.
Key Innovation: Following the analysis of various contributing error factors, this study focuses on the positioning errors of HIRAS by constructing a geometric positioning and correction model from each field of view (FOV) to the ground.
164. A machine learning framework for predicting agricultural terrace suitability using positive and unlabelled datasets: an application for the Troodos Mountains, Cyprus
Core Problem: Terraces are a defining feature of Mediterranean mountain landscapes, enabling agriculture on steep slopes while providing multiple ecosystem services.
Key Innovation: The developed PU-classifier was evaluated under the selected completely-at-random (SCAR) assumption, achieving a Recall of 84.6%, Precision of 81.5% and an F1 score of 83%.
165. Stratified and adaptive regression for NDVI-based land surface temperature downscaling over agricultural fields
Core Problem: Improving the spatial resolution of thermal imagery is essential for agricultural field management, especially in developing countries where fields are fragmented and heterogeneous.
Key Innovation: This presents an improvement of an average of 33 % compared to the two classical methods DisTrad and TsHARP (RMSE in the range 4.18-6.83 °C), which showed similar accuracy.
166. Federated learning in forest resource modelling and monitoring: Bridging data confidentiality and collaborative research
Core Problem: The availability of reliable ground-truth data is one of the main bottlenecks for improving high-resolution forest attribute maps from Earth observation data.
Key Innovation: FL models achieved predictive performances comparable to the traditional models, which proofs the effectiveness of the proposed approach.
167. Baseline framework for inferring POI floors to enhance 3D urban vertical analytics: A comparative study in Shanghai
Core Problem: However, most open POI datasets lack vertical attributes, causing incomplete semantic representations.
Key Innovation: The results demonstrate the feasibility of inferring POI floor levels from partially labeled open data using lightweight features, which can support vertical semantic enrichment for 3D urban analytics and serve as an updateable semantic enrichment module for urban digital twins, when new POI and building data become available.
168. A double-temperature-difference framework for anthropogenic heat flux estimation accounting for surface energy balance perturbations
Core Problem: The proposed DTD framework provides a physically constrained and EO-supported approach for city-scale AHF mapping and can support urban climate modeling and thermal-environment assessment.
Key Innovation: In this study, a Double Temperature Difference (DTD) framework was developed to estimate AHF from land surface temperature and near-surface air temperature responses.
169. Automatic design of efficient SAR ship detection networks for large-scale scenes via neural architecture search
Core Problem: Traditional deep learning-based Synthetic Aperture Radar (SAR) ship detection networks rely on expert manual design, making it difficult to adapt to the diversity of targets and environments.
Key Innovation: On the SSDD and HRSID datasets, the proposed method achieves search times of 1.981 h and 11.449 h, respectively, with mAP50 detection accuracies of 98.2% and 92.3%, respectively.
170. UVSENet: A Semantic-Embedded dual branch fusion network for urban villages mapping
Core Problem: However, traditional optical remote sensing imagery is inherently limited by insufficient single modality representation capability, while multimodal data fusion is often constrained by data heterogeneity and cross modal semantic inconsistency.
Key Innovation: This design further facilitates cross regional knowledge generalization, thereby improving the recognition of complex spatial patterns characteristic of UVs.
171. Identifying multi-scale human flow clusters via spatiotemporal network flow L function
Core Problem: Understanding spatiotemporal clustering patterns of human flows is critical for interpreting mobility dynamics and addressing urban issues.
Key Innovation: The study introduces the linear reference system to explicitly represent human flows on linear road networks, and proposes the Spatiotemporal Network flow K-function (STNflowK) and its variation (STNflowL function) for multi-scale human flow clustering.
172. GSR-Track: A Geo-Semantic Regularized Sparse tracker for satellite videos
Core Problem: Satellite video object tracking (SVOT) faces general challenges such as tiny target sizes, dense distractors, and complex environmental occlusions.
Key Innovation: To systematically address these architectural limitations, we propose the Geo-Semantic Regularized Sparse Tracker (GSR-Track).
173. Reconstructing history: 3D high-resolution GPR analysis based on SAR discoveries in the ancient Siniyah Island of Umm Al-Quwain, United Arab Emirates
Core Problem: SAR tomography provided intermediate-scale subsurface assessment, reducing spatial uncertainty prior to field investigation.
Key Innovation: Here, we present for the first time on the island, A high-resolution Ground Penetrating Radar (GPR) survey was carried out in three priority areas delineated through an integrated Synthetic Aperture Radar (SAR) analysis and machine-learning-based spatial screening.
174. Infinite details: A cartography-informed parametric method for constructing realistic 3D natural scenes
Core Problem: However, achieving realism in natural scenes is challenging because of thecomplexity of microscale details influenced by natural disturbance events such as weathering, aging, and decay.
Key Innovation: The results demonstrate the adaptability and effectiveness of the proposed method in scenes comprising various types of entities and details.
175. SignSpotter: scene text-informed cross-modal semantic consistency verification for street view images and POI data
Core Problem: Ensuring semantic consistency between street view images and Points of Interest (POI) is critical for reliable geographic entity validation in complex urban scenes.
Key Innovation: To address challenges such as visual occlusion, scale variation, and ambiguous scene text, this paper proposes SignSpotter, a novel multi-scale vision-language framework for entity-level consistency verification.
176. The role of atmospheric correction errors in the underestimation of satellite-derived chlorophyll-a concentration in the Southern Ocean
Core Problem: While this bias is often attributed to the region’s unique bio-optical properties, the potential contribution of atmospheric correction errors has been comparatively underexplored.
Key Innovation: Our results show that when in-situ Rrs were used as inputs, the mean absolute percent difference (MAPD) values between measured and estimated CHL by OC2, OC3, and OCI were 24.4%, 23.0%, and 30.2%, respectively.
177. OpenPineDet: An open-set detection framework for the detection of dead pine wood in forests
Core Problem: However, these approaches typically assume that both the training and test datasets contain only instances from a single, known class (i.e., DPW).
Key Innovation: OPDet integrates two novel modules: (1) a Bottleneck Channel Attention Network (BNCA) that leverages the Efficient Channel Attention (ECA) mechanism to enhance the model’s focus on DPW targets; and (2) a Dual Contrastive Learning (DCL) module, which combines Instance-Level Contrastive Learning (ICL) with Focal Loss (FL) to improve detection accuracy of hard-to-classify DPW instances.
178. Spatially explicit predictions of land cover change using the Density-Functional Fluctuation Theory (DFFT) approach
Core Problem: Compared with commonly used approaches such as CA-Markov, DFFT is highly parsimonious and scalable, making it particularly suitable for large domains, data-scarce settings, and scenario-based forecasting.
Key Innovation: The study introduces Density-Functional Fluctuation Theory (DFFT) as an alternative, low-data and computationally efficient approach for spatially explicit LCC prediction.
179. DBE-Net: a dual-branch bounding box-embedding network for individual tree segmentation from LiDAR
Core Problem: However, the state-of-the-art reported works struggle to accurately segment individual trees in scenarios characterized by canopy overlap and low-density point clouds.
Key Innovation: The core novelty of DBE-Net lies in its dual-branch architecture that fuses proposal-based and clustering-based instance segmentation paradigms, which effectively alleviates the reliance on a single prediction branch and improves both accuracy and efficiency.
180. Geometry-aware mapping with MI-IWNet reveals topographic modulation effects in internal wave evolution
Core Problem: Accurate mapping of oceanic internal waves (IWs) is essential for understanding ocean mixing and energy transfer, yet remains challenging due to weak and transient sea-surface signatures and heterogeneous sensor limitations.
Key Innovation: In this study, we propose a geometry-aware deep learning framework, MI-IWNet, which incorporates a direction-aware encoder design (including Direction-Aware Strip Convolution, DASC) to represent the anisotropic and curvilinear morphology of IW crests in multi-sensor satellite imagery; we further develop a graph-based Structure-Aware Wave Packet Aggregation algorithm to assemble pixel-level predictions into object-level Oriented Bounding Boxes (OBBs) for robust geophysical analysis.
181. A novel framework for the selection of drifter deployment sites to optimize satellite particle backscatter validation
Core Problem: However, the paucity of in situ multi-band b b p data hinders efforts to quantify uncertainties in satellite b b p and its derived products.
Key Innovation: The high sampling frequency, combined with a Lagrangian approach, enables it to overpass numerous pixels in a single day, thus providing large in situ datasets for validation of b b p , not achievable by other in situ platforms.
182. Urban-TreeSeg: A CRUNet-based deep learning framework for urban tree crown segmentation from archived aerial imagery
Core Problem: However, accurate delineation of individual tree crowns - a prerequisite for reliable biodiversity assessment - remains challenging due to factors such as overlapping canopies, heterogeneous backgrounds, limited spectral information, and diverse tree crown sizes.
Key Innovation: The framework integrates three key components: (1) a CBAM Residual UNet (CRUNet) architecture for enhanced multiscale feature extraction, (2) a classification branch to improve separation of tree crowns from urban backgrounds, and (3) star-convex polygon representation to better capture irregular crown shapes.
183. Local shadow compensation considering feature differences for UAV images
Core Problem: Shadows become evident in UAV images, which will cause contrast reducing, loss of details and texture blurring, resulting in the decrease of image quality and affecting the accurate identification of ground object information.
Key Innovation: Therefore, we propose a novel method of local shadow compensation considering feature differences (LSCFD) for UAV images, which resolves the overlooked spectral heterogeneity inside shadow regions through feature-aware segmentation and statistical analysis.
184. Crop multiclass classification for large-scale crop mapping using instance- and feature-domain integrated transfer learning
Core Problem: Although previous research has successfully achieved large-scale crop mapping using time-series features derived from remote sensing imagery, classification schemes integrating crop-specific rhythms and phenological knowledge remain insufficient under target-label-free conditions in the study area.
Key Innovation: Integrating phenological and crop features notably enhanced model performance, increasing the learning curve score by 0.11; (2) The CMSTA method achieved superior crop mapping results within the YRD, with an Overall Accuracy (OA) and Weighted F1-score of 0.90.
185. Ground-penetrating radar signature of Albic Podzols in the boreal Forest zone: Key characteristics and its implications for digital soil mapping
Core Problem: Ground-penetrating radar interpretation of forest soils is ambiguous because weak horizons and site-specific electromagnetic variability obscure diagnostic boundaries.
Key Innovation: Seventeen soil profiles and two field sites establish depth-dependent dielectric and conductivity signatures for Albic Podzols and quantify links between permittivity, organic matter and structural units.
186. Impacts of herbaceous plant roots on the engineering properties and erodibility assessment of the ming great wall earthen sites in the alpine and humid-cold regions
Core Problem: Although vegetation can mitigate surface erosion, the mechanisms by which herbaceous roots regulate vertical soil erodibility in earthen heritage sites remain unclear.
Key Innovation: An improved Comprehensive Soil Erodibility Index (CSEI) was developed using a nonlinear Sigmoid function by integrating mean weight diameter (MWD), soil disintegration rate (SDR), saturated hydraulic conductivity (SHC), cohesion (Coh), and the soil erodibility factor (K).
187. Coupled effects of organic matter and salinity on thermal conductivity of frozen sands
Core Problem: Organic matter and salinity, which are common in permafrost, alter the thermo-hydraulic behavior of soils in cold regions; therefore, improved characterization of the combined influence of these constituents on thermal properties has become a pressing need for accurate prediction of subsurface processes.
Key Innovation: The outcomes of this study help establish linkages between soil composition and changing environmental conditions and support improved prediction of thermo-hydraulic interactions in permafrost and seasonally frozen soils.
188. Coupling analysis of mechanical-electrical characteristic of unsaturated frozen sulfate red clay under dynamic loading
Core Problem: As transportation infrastructure extends into high-altitude and high-latitude regions with complex environmental conditions, the safety of subgrades and underground structures in permafrost areas has attracted increasing attention.
Key Innovation: The results show that under cyclic loading, electrical resistance gradually decreases and the absolute value of reactance progressively diminishes, indicating continuous optimization of conductive pathways and weakening of polarization effects.
189. Research on local thermal non-equilibrium boundary of crushed rock embankment
Core Problem: Accurately describing the heat transfer mechanism at the crushed rock surface plays a critical role in predicting the long-term service performance of crushed rock embankments.
Key Innovation: Based on the surface energy balance equation, this study investigates the local thermal non-equilibrium (LTNE) heat transfer mechanism that characterizes the thermal boundary of the crushed rock skeleton under the coupled interaction of radiation, convection, and conduction at the surface of the crushed rock layer, and develops a new calculation method for the thermal boundary.
190. Visual experimental investigation on dynamic propagation mechanisms of multi-fractures in layered media
Core Problem: A mechanistic understanding of the dynamic initiation and propagation of multi-cluster hydraulic fractures in layered reservoirs is essential for optimizing fracture-height growth and plugging strategies during volumetric fracturing.
Key Innovation: In this study, a visual hydraulic fracturing system was established using transparent polymethyl methacrylate (PMMA) specimens, and a high-speed camera was employed to capture the transient evolution of fracture morphology in real time.
191. Spatiotemporal dynamics of capillary rise in shallow groundwater table areas revealed by spatial TDR
Core Problem: Accurate quantification of soil water content (θ, cm3 cm-3) dynamics and capillary rise in shallow groundwater table areas remains a major challenge due to the limitations of conventional discrete point-scale measurements of θ.
Key Innovation: The study presents an integrated monitoring sensor system based on spatial time domain reflectometry (TDR) for retrieving high-resolution θ profiles along transmission lines.
192. Vegetation-hydrology interactions modulate porewater salinity dynamics in coastal salt marshes
Core Problem: Although water-salt dynamics under tidal and climatic variations are well studied, the dynamic role of marsh vegetation in regulating porewater salinity remains insufficiently understood.
Key Innovation: The study integrates field monitoring of soil bulk electrical conductivity (EC) with controlled experiments to assess how upper intertidal vegetation influences bulk-EC-inferred porewater salinity dynamics, using Spartina alterniflora as a representative marsh species.
193. Modeling fracture growth in porous media considering solid dissolution with the assumed enhanced strain method
Core Problem: The study presents an assumed enhanced strain (AES) finite element framework for simulating fracture growth in porous media, accounting for weakening effect induced by solid grain dissolution.
Key Innovation: The numerical results demonstrate that the proposed framework can provide effective simulations and therefore help better understand fracture growth with solid dissolution in poroelastic media.
194. Development and application of a creep model of rock mass based on memory-dependent derivatives
Core Problem: To overcome the difficulty of accurately characterizing the stress-history dependence and nonlinear acceleration of the creep of rock mass, a new creep model of rock mass (HAMD) was developed based on the memory-dependent derivative (MDD).
Key Innovation: The model was constructed by connecting an elastic element, Abel dashpot, and strain-triggered memory-dependent element in series.
195. Analytical solution of coupled transient liquid flow, vapor flow, and heat transfer in unsaturated soils
Core Problem: Coupled liquid, vapor, and heat transport processes play a central role in the behavior of unsaturated soils, yet their strong nonlinear interactions have made analytical solutions elusive.
Key Innovation: The study introduces a new analytical framework capable of solving the fully coupled transient liquid-vapor-heat system governing moisture and thermal redistribution in unsaturated porous media.
196. Seismic soil-structure interaction and deformation mechanism of pile-supported frame integral abutment bridges
Core Problem: However, the seismic soil structure interaction mechanisms and deformation patterns of pile-supported IAB are not well understood.
Key Innovation: A numerical model is established to investigate two critical instants associated with the onset of abutment failure and is compared with the experimental results.
197. Strength-deformation and critical state of coarse-grained soil considering particle shape and particle breakage
Core Problem: In this study, a series of consolidated drained triaxial tests were performed to evaluate the effects of shape and breakage on the mechanical and critical state behavior of coarse-grained soils.
Key Innovation: Coarse-grained soils are extensively employed as primary construction materials in civil infrastructure, including high earth-rock dams and railway embankments, where long-term serviceability is of paramount importance.
198. Thermo-hydro-mechanical coupling effects on cyclic dynamic behavior of saturated soft soil
Core Problem: Cyclic thermo-hydro-mechanical (THM) behavior of saturated soft soils remains insufficiently quantified, particularly regarding long-term deformation mechanisms.
Key Innovation: Temperature-controlled dynamic triaxial tests were systematically performed to evaluate thermo-mechanical responses in saturated soft soils.
199. Investigating strain localization at cracked concrete-sandstone specimens using optimized fractal theory-based image thresholding segmentation algorithm
Core Problem: The stability of concrete-rock interfaces is a critical issue in underground engineering.
Key Innovation: Furthermore, the proposed method detected a three-phase evolution in fractal dimension before failure, which follows an initial increase, a stable period, and a final rapid rise.
200. Transverse vibration frequency and the key influence factors of fully grouted rock bolts in layered rock mass
Core Problem: Health monitoring of fully grouted rock bolts (FGRBs) remains challenging in geotechnical engineering, despite their widespread use as primary support elements in fractured rock masses.
Key Innovation: The study established a governing equation for the transverse vibration and derived an analytical expression for the FGRB’s transverse vibration frequency (f) based on beam transverse vibration theory.