Search Results (777)

Rapid and reliable assessment of structural damage following disasters is critical for prioritizing rescue operations. In this study, we present a unified deep learning framework for building damage assessment from satellite imagery. The proposed approach integrates segmentation-driven feature construction, morphological processing, cost-sensitive learning, and cross-disaster evaluation to support robust performance under limited and imbalanced data conditions. The framework combines an adapted U-Net for building localization with a hybrid convolutional neural network (CNN)-deep neural network (DNN) classifier for damage-level prediction and evaluates transferability across disaster events, geographic regions, and sensing conditions. The proposed method is evaluated on selected events from the xView2 Building Damage Assessment (xBD) and BRIGHT datasets, using optical imagery from xBD and pre-disaster optical and post-disaster Synthetic Aperture Radar (SAR) imagery from BRIGHT. Despite the limited and highly imbalanced event-specific samples, the framework achieves a mean cross-validation macro-F1 score of 70% and a maximum fold-level score of 77% on the Mexico earthquake subset of xBD and up to 98% on earthquake-related events in BRIGHT. Cross-validation characterizes performance variability across source-image-grouped data partitions, while cross-disaster evaluation reveals event-dependent transferability and provides a preliminary indication that structural domain similarity may be related to transfer performance. Although the evaluation is constrained by data availability, the results indicate that lightweight, cost-sensitive deep learning frameworks may support auxiliary post-disaster screening and decision support in resource-constrained scenarios. This study highlights both the potential and the remaining challenges of deploying artificial intelligence (AI) for rapid post-disaster assessment. Full article

Reservoir impoundment can reactivate pre-existing landslides and reorganize slope topography, thereby changing seepage conditions and subsequent deformation. However, crack mapping, geomorphic interpretation, and hydrodynamic diagnosis are still often treated as separate tasks. This study investigates the Hemenkou (HMK) landslide in the Wudongde Reservoir area, China, using multi-scale space–air–ground observations, including multi-temporal optical satellite images, unmanned aerial vehicle (UAV) photogrammetry, pyramid scene parsing network (PSPNet)-based crack segmentation, global navigation satellite system (GNSS) monitoring, and convergent cross mapping (CCM). The remote sensing record shows a progressive damage sequence: cracks were mainly restricted to the upper source area in 2012, crown cracking intensified and propagated downslope by December 2020, and the UAV survey of 10 June 2024 revealed a mature tension-crack network concentrated in Zone II. ResNet-50-PSPNet achieved the best crack-extraction performance among the tested models, with Precision = 0.9120, Recall = 0.9041, F1 = 0.9081, and IoU = 0.8316. The mapped cracks are dominated by short, narrow, northeast–southwest-oriented tension cracks. GNSS monitoring reveals strong spatial heterogeneity, with stepwise deformation concentrated in Zone II. CCM provides strong directional evidence for the influence of reservoir water-level fluctuation on Zone II deformation, whereas the weaker rainfall signal is consistent with a secondary reinforcing role. The apparent increase in the rainfall-related CCM signal from 2021 to 2023 is consistent with progressive crack expansion and potentially enhanced hydraulic connectivity in Zone II. Taken together, these observations support the interpretation that post-deformation topography, particularly the tension-crack network and disturbed toe, may organise preferential seepage pathways and increase the sensitivity of the landslide to reservoir drawdown. The study provides an integrated remote sensing and monitoring framework for process-based interpretation of reservoir landslides. Full article

In slow-time colorized subaperture image (CSI), anisotropic targets that reflect strongly when viewed from specific angles appear in vivid colors, which makes them stand out against isotropic background that reflects energy uniformly across all angles. It leads to more accurate annotation labels for ships, vehicles, and airplanes in SAR images and better SAR automatic target detection (ATD) performance. Unfortunately, although many port-related CSI products collected by satellite-borne SAR systems are released for free public access and could be leveraged for ship detection research, those that could support vehicle and airplane detection are rare. To investigate performance improvement in deep learning-based SAR ATD that could be brought by colored SAR images, three novel SAR-ATD frameworks are proposed for ship, vehicle, and aircraft detection, respectively. (1) Context-guided ensemble learning (CGEL) is proposed for ship detection, where state-of-the-art high-resolution colorized spotlight SAR images are exploited to enhance the visual features of ships and reduce false alarms, while the potential ship berthing/docking areas are delimited with adaptive intensity shading (AIS). (2) Context-driven SAR image recoloring and enhancement mechanism (CD-SAR-REM) is proposed to generate a context-driven color-enhanced version of the original SAR image based on AIS so that potential parking regions are highlighted. (3) Color feature-aided aircraft detection. In case that CSI products are unavailable, pseudo-color SAR images are generated based on phase congruency and the contextual information extracted by the segmentation module is used to refine the initial predictions generated by the core detection network. Experimental results show that the performance of the proposed context-driven ship, vehicle, and aircraft detection methods based on colored SAR images are superior to many state-of-the-art SAR ATD models. Full article

In Deep Learning Remote Sensing, data quantity is rarely the limiting factor. A single high-resolution satellite image can yield thousands of training patches. What determines model performance, yet remains largely overlooked, is the quality of those patches. To date, the choice of sampling method has rarely been treated as a methodological decision. Conventional approaches, namely sliding-window and random sampling, introduce two compounding data-quality problems: severe class imbalance caused by the overproduction of background-only patches and negative learning arising from incomplete annotations, where unlabeled objects are implicitly treated as negative examples during training. To address these limitations at the data construction stage, we propose object-centric patch sampling, a model-independent strategy that anchors each training patch to the geometric centroid of an annotated object. This design ensures that every object-anchored patch contains at least one target instance and substantially reduces exposure to unlabeled regions that generate false-negative supervision signals; only a small, deliberately controlled proportion of background-only patches is retained to preserve contextual variety without reinstating background dominance. The method is evaluated on three heterogeneous remote sensing datasets spanning satellite (Sentinel-2, 10 m), aerial (NAIP, 1 m), and UAV (0.25 m) imagery, covering cotton field segmentation, rural building extraction, and water body delineation, respectively. Using a U-Net architecture under identical training conditions, the proposed approach achieves IoU scores of 0.929, 0.896, and 0.912 on the three datasets, respectively, outperforming sliding-window sampling by up to 19.6 percentage points in IoU and consistently delivering higher F1-scores across all experimental configurations. Evaluation under DeepLabV3+ gives a mean IoU of 0.9504 and a mean F1-score of 0.9772 in a multi-class segmentation task, indicating that the gains are not specific to a single architecture. Unlike model-level solutions such as focal loss or class reweighting, the proposed method improves training data quality at its source and integrates seamlessly into any deep learning pipeline without architectural modifications. Full article

Cross-view geo-localization (CVGL) is a critical technology used in unmanned aerial vehicles (UAVs) and widely applied in navigation and target localization tasks. However, owing to the extreme perspective disparity between UAV oblique views and satellite vertical views, CVGL still involves significant challenges, including geometric distortion caused by viewpoint differences, drastic appearance inconsistencies, and the difficulty in bridging semantic gaps between heterogeneous data. To address these issues, we propose a novel CVGL method named dynamic-feature collaborative optimization and semantic-alignment network (DFSA), designed to extract robust feature representations and achieve fine-grained alignment. Specifically, the DFSA employs a residual-based vision transformer as the backbone to capture global context while alleviating the training instability and feature collapse often associated with standard transformers. To bridge the semantic gap between global and local features, we design a feature optimization module comprising a local feature enhancer and a global feature aggregator. This module establishes a closed-loop collaborative system that facilitates top-down semantic guidance and bottom-up detail feedback. Furthermore, we introduce a semantic segmentation and alignment module that adaptively partitions images into semantic regions based on feature response distributions, shifting the matching granularity from the global level to the semantic region level to effectively overcome feature mismatches caused by positional offsets and scale variations. Extensive experiments conducted on the University-1652 and SUES-200 datasets demonstrate the superior image retrieval performance of the proposed DFSA. Specifically, DFSA achieves a Recall@1 of 94.87% and an Average Precision (AP) of 95.32% on the University-1652 dataset and maintains highly competitive Recall@1 performances between 96.83% and 99.25% across various altitudes on the SUES-200 dataset. These results validate the model’s effectiveness in handling extreme viewpoint changes for UAV-based cross-view image retrieval tasks. Full article

Accurate and up-to-date knowledge of land use and land cover represents one of the central challenges in spatial planning and landscape sciences. In this context, the present work introduces GANCIU (Geospatial Analysis with Neural Classification and Image Understanding), an original hybrid pipeline for the automatic extraction of man-made infrastructure from high-resolution satellite imagery. The primary methodological contribution lies in the sequential integration of four technologically heterogeneous components: a per-pixel Random Forest classifier, a guided image modulation step, edge detection via the Mumford–Shah variational functional solved through the Ambrosio–Tortorelli approximation, and final object delineation via the Segment Anything Model (SAM). Each component does not operate independently but conditions and informs the next: The RF probability map guides the modulation, which in turn directs the sensitivity of the variational step exclusively towards regions of interest; the AT edges provide spatial prompts to SAM, for which its masks are finally filtered by the RF probability in an adaptive manner through a Gaussian Mixture Model. This progressive conditioning scheme constitutes the architectural core of GANCIU and distinguishes it from approaches that combine classification and segmentation in parallel or in purely sequential fashion with each stage conditioning the next but without any reverse correction between them. The Random Forest classifier was trained on 44 manually annotated scenes, geographically disjoint from the twelve independent scenes used for quantitative validation. This validation, based on an instance matching protocol (precision, recall, F1 score, and IoU), confirms the contribution of the full pipeline over a Random-Forest-only baseline: Pooled false positives fall by close to two orders of magnitude (from 8320 to 209), while true positives rise nearly twentyfold (from 5 to 95), with a mean IoU of 0.742 ± 0.060 on correctly matched objects. Notably, the entire pipeline—including SAM-based segmentation—runs end-to-end on a modest, GPU-free consumer laptop (four logical CPU cores, under 16 GB RAM), demonstrating that competitive infrastructure-extraction performance does not require specialised computing hardware. Full article

High-resolution and regularly updatable land cover maps are essential for local-scale environmental monitoring, water resource management, and territorial governance, yet existing global and regional products fail to provide the spatial detail and thematic richness required for operational applications in complex Mediterranean landscapes. This study presents an operational workflow for high-resolution LULC mapping and its application to Sardinia for the reference year 2020, developed within the Sardinia Land Cover Mapping Project in collaboration with the Agenzia del Distretto Idrografico della Sardegna (ADIS). The workflow integrates multi-temporal SAR and multispectral satellite imagery with high-resolution ancillary geospatial vector datasets through a semi-automatic pipeline combining hierarchical cascade pixel-based classification, multi-resolution image segmentation, geometric overlay of infrastructure vector layers, and an iterative accuracy-driven reclassification cycle. The classification combines automated rule-based procedures, semi-automatic threshold-based methods, and expert photo-interpretation to address the high thematic and spatial complexity of the Sardinian landscape. The resulting map comprises 35 land cover classes at the third and selected fourth CORINE levels, with a minimum mapping unit of 400 m² and an overall weighted accuracy of 82.4%. Designed as a dynamic product updatable on an annual basis, it represents an operational tool for local environmental governance, spatial planning, and resource management. Full article

Precise analysis of crop fields is essential for agricultural management. Various remote sensing tasks comprise the characterization of agricultural lands. The automatic extraction of boundaries and the crop type semantic segmentation tasks, both benefiting from satellite image time-series analysis, are crucial for continuous field monitoring. Recent models explore the multitask setting for boundary detection, but they do not incorporate class information that is crucial when several kinds of agricultural crops co-exist. Inspired by the idea that crop type segmentation and boundary extraction are closely connected, this study exploits a multitask learning framework for both agricultural segmentation and boundary detection tasks. The model uses a shared encoder with 3D convolutional blocks and task-specific decoders with cross-task feature exchange to simultaneously solve the two main tasks of crop type segmentation and boundary extraction. The model also leverages on estimating the distance to the closest border as an auxiliary task to improve training. The model effectively shares features across tasks and solves them simultaneously, achieving detailed crop-field characterization. To validate the method performance and examine inter-task relationships, two datasets composed of time series of Sentinel-2 images acquired from two agricultural areas in Austria, spanning two different years, were used. Full article

A UAV operating without GNSS faces unbounded inertial drift. A layered navigation architecture is evaluated in which terrain contour matching (TERCOM) provides periodic position corrections and satellite-image scene matching adds a condition-dependent precision layer. The architecture is examined through a single-trajectory simulation over a 1° × 1° ASTER GDEM V2 tile (N31E081, Tibetan Plateau, 4555–6468 m elevation, 16.1 mean slope) representing a one-hour flight (127 km, 35.2 m/s). The simulation models GNSS loss with idealised sensor behaviour: IMU error is described by a Gauss–Markov model without temperature dependence, and the radar altimeter is represented with additive Gaussian noise. Under these conditions, TERCOM reduced RMS position error from 1467 m to 317 m (78.4% reduction); with ideal noise-free scene-matching registration added, RMS further decreased to 103 m (a best-case estimate). The idealised Cramér–Rao lower bound already incorporates the 5 m radar-altimeter and 20 m DEM noise terms (it is therefore not a noise-free value) at the flight mean slope of 16.1°; averaging this local bound over the full trajectory—where near-flat segments inflate it—gives the tile-averaged CRLB of ≈150 m. The remaining gap between the realised TERCOM RMS (317 m) and this realistic bound is attributed to residual INS drift during profile collection, DEM interpolation error, and low-entropy terrain segments; a quantitative decomposition of these factors is provided in this paper. Results are based on a single noise realisation and a single trajectory; they characterise the specific simulation scenario rather than the architecture’s general performance. The altitude-error decomposition argument—that TERCOM’s sensitivity depends primarily on short-term dynamic altitude drift rather than the accumulated systematic error—is developed specifically for the normalised cross-correlation (NCC) metric and requires mean-centring of the terrain profile for generalisation to other correlation metrics. Full article

Urban tree mapping is necessary in environmental sustainability and climate change mitigation, and it depends heavily on the individual tree recognition and canopy segmentation to analyze city green cover. This systematic review discusses recent developments in the 2014–2026 mapping of these trees with the use of UAVs and high-resolution satellite imagery. Our preliminary selection of 4148 records reduced to 101 eligible publications following a systematic screening and synthesis of the records, assessed the efficiency of deep learning models such as Convolutional Neural Networks and Vision Transformers in processing various source images. We also compare object detection and semantic segmentation to see which one is more competent to deal with typical urban challenges, including overlapped canopies and building shadows. According to the reviewed studies, UAV-based models generally achieve higher spatial accuracy than satellite-based approaches for individual tree detection and crown delineation, with reported average Intersection over Union (IoU) values of approximately 70–75%, whereas satellite imagery provides superior spatial coverage for large-scale urban forest monitoring. Lastly, we present a research roadmap to address the existing weaknesses such as geographic bias, which propels the research direction towards multimodal data fusion and Foundation Models to sustain consistent, large-scale urban forest monitoring. Full article

Accurate greenspace quantification underpins sustainable urban greening management. Remote sensing (RS)-derived green quantity is a core urban planning indicator, yet its capability to reflect actual urban green supply lacks systematic verification, inevitably affecting planning formulation and decision-making. Taking Shanghai as the study area, this study integrates remote sensing and Baidu Street View (BSV data) to compare the two-dimensional planar green coverage derived from satellite imagery with pedestrian-level perceived visual green coverage. The Mask2Former model was adopted for high-precision semantic segmentation of BSV images to extract vegetation, building, sky and hard pavement proportions. Geographically Weighted Regression (GWR), hotspot analysis and transition mapping were applied to identify divergent regions, while the XGBoost-SHAP framework was employed to explore deviation mechanisms. The results reveal distinct spatial pattern differences between the two green quantity datasets. RS-derived green quantity exhibits strip-like agglomeration, whereas BSV-perceived green quantity is more fragmented, with a correlation coefficient of only 0.240. Single RS quantification fails to reflect street-level green supply. Divergent areas are classified into accurate, overestimated and underestimated zones. RS underestimates green quantity in central urban areas and overestimates that in northwest suburbs. BSV-based sky ratio, building ratio, hardscape ratio and Road 1 density are core influencing factors with obvious nonlinear threshold effects. This study clarifies the quantitative deviation patterns and mechanisms between RS and BSV green quantity, providing scientific support for precise urban green planning and sustainable perceived visual green construction. Practically, the dual RS–street view assessment framework proposed in this paper can be embedded into routine urban green infrastructure auditing, help planners distinguish systematically overestimated suburban green belts and underestimated central urban micro-green spaces, and deliver targeted optimization strategies for vertical greening, street tree renovation and pocket park construction under high-quality urban renewal demands. Full article

Fine-grained mapping of riparian vegetation is important for ecological monitoring, invasive species control, and ecosystem restoration. However, riparian plant communities often exhibit fragmented patches, broad transition zones and high visual similarity among classes, making stable species-level segmentation difficult from conventional satellite imagery or lower-resolution UAV imagery. To address this gap, we present WetVeg-2mm, an ultra-high-resolution UAV dataset for fine-grained riparian vegetation semantic segmentation. Built from UAV surveys over a representative riparian section of the Jiuzhou River in Guangxi, China, the dataset provides 2054 image chips (1024 × 1024) with pixel-level annotations at 2 mm ground sampling distance. It contains 17 semantic classes in total, including 14 representative wetland plant classes, such as Colocasia, Eichhornia and Phragmites, together with water, bareland and background. Five baseline models, namely U-Net, Attention U-Net, DeepLabV3+, PSPNet and SegFormer, were evaluated using per-class IoU, mIoU, mDice, PA, Precision and Recall. Across all evaluated baseline settings, SegFormer with ImageNet pretraining achieved the best overall performance, with 76.23% mIoU, 86.01% mDice, 85.17% PA, 87.30% Recall and 85.42% Precision on the test set. Overall, WetVeg-2mm provides a reproducible and challenging benchmark for fine-grained riparian vegetation semantic segmentation. Full article

Reliable assessment of shorelines extracted from medium-resolution satellite imagery requires independent high-resolution reference data and statistical methods that account for spatial dependence. This study compared three conventional analyst-assisted shoreline-extraction workflows—histogram thresholding, band ratio, and the Normalised Difference Water Index (NDWI)—at Coronation Park, Lake Ontario, Canada, using Sentinel-2 Level-2A imagery. A manually digitised shoreline derived from a UAV-based orthomosaic acquired approximately 27 h before the Sentinel-2 scene served as the independent reference. The UAV-based reference and each Sentinel-2-derived shoreline were divided into 31 ordered segments. For each Sentinel-2-derived segment midpoint, the shortest planar Euclidean distance to the nearest UAV-based reference midpoint was calculated and used to derive mean absolute error (MAE) and root mean square error (RMSE). Residual spatial autocorrelation was assessed using Moran’s I with 9999 permutations. Because the paired differences departed from normality, the Friedman test was treated as the primary overall comparison, while contiguous spatial-block permutation tests across block sizes of two to eight shoreline locations assessed robustness to local spatial dependence. NDWI achieved the highest positional agreement (MAE = 5.645 m; RMSE = 6.429 m), followed by band ratio (MAE = 14.303 m; RMSE = 14.797 m) and histogram thresholding (MAE = 26.167 m; RMSE = 26.910 m). Significant positive residual spatial autocorrelation was identified for all three methods (Moran’s I = 0.587–0.832, all p < 0.001). The Friedman test confirmed a significant extraction-method effect, χ²(2) = 49.226, p < 0.001, Kendall’s W = 0.794, and the effect remained significant across all tested spatial-block sizes, with empirical p-values ranging from 0.000007 to 0.004630. Among the three conventional methods tested at this large-lake site, NDWI provided the highest positional agreement and therefore offers a defensible baseline for evaluating future Sentinel-2 image-enhancement approaches. Full article

In this study, changes in the surface area of Lake Burdur during the 2015–2025 period and the spatial direction of the associated shrinkage were examined using Sentinel-2 Level-2A satellite images. A total of 111 satellite images, each representing a monthly period, were analyzed using a fixed study window and a lake vicinity mask; a three-cluster unsupervised K-means segmentation method was applied to separate the water surface from bare/drained areas and vegetation classes. The resulting binary water masks were used to convert the lake surface area to km² on a pixel-by-pixel basis, and the geometric center of the lake mass was calculated for each observation date. The unique aspect of this study is that it evaluates lake shrinkage not only through a decrease in surface area but also as a directional spatial process via the movement of the centroid center. In this context, the cumulative displacement was decomposed into X/West and Y/South components using the initial centroid point as a reference; OLS-based linear and logarithmic trend models were established for both directions. Model performances were compared using a 15-fold Monte Carlo cross-validation approach with R², adjusted R², NSE, KGE, MAE, MAPE, MSE, and RMSE metrics; additionally, the statistical significance of model differences was assessed using the Wilcoxon signed-rank test. The findings indicate that the logarithmic model yields more balanced and reliable results in the X/West direction, while the linear model does so in the Y/South direction. Based on this model structure, spatial projections were generated for the 2026–2035 period, and it was observed that the projection bands remained stable despite Monte Carlo-based coefficient uncertainty. In conclusion, the study demonstrates that lake drawdowns in semi-arid closed basins can be monitored in a more interpretable and statistically robust manner using Sentinel-2-based segmentation, centroid kinematics, and cross-validated trend modeling. Full article

The paper addresses the problem of maritime navigation in coastal areas without access to satellite positioning data. The proposed approach relies on visual information from a monocular camera supported by map-derived coastal information. Although the operational concept assumes the use of publicly available cartographic data, the method’s preparation and evaluation require additional processing steps, including high-resolution aerial imagery, GIS-based extraction of coastal features, semantic segmentation, and trained convolutional neural network models. The problem discussed in the paper concerns, for example, Autonomous Underwater Vehicles that seek to reduce underwater dead-reckoning navigation error by surfacing and using information about what is visible around them, in a way similar to how a human would. To solve the above problem, a system was proposed that compares the camera’s representation of the observed coastline with the map representation of the area where the vehicle is most likely located. The system was validated using real-world data. The tests revealed that the information contained in a flat map is insufficient for accurate position estimation. Accuracy is also significantly affected by errors in the semantic segmentation used to extract land features from camera images, as well as by potential errors in the camera viewing angle. The achieved accuracies are sufficient for navigation away from land, but when operating close to land, the proposed system appears significantly insufficient. The paper specifies the system and reports the results. Full article