Remote Sensing

27 pages, 8211 KB

Open AccessArticle

Dual-Level Spatial–Frequency Collaborative Detector for Oriented Object Detection in Remote Sensing Images

by Xuehuai Shi, Jingru Sun, Kun Yu, Zhihui Wei and Shangdong Zheng

Remote Sens. 2026, 18(16), 2845; https://doi.org/10.3390/rs18162845 - 21 Aug 2026

Viewed by 458

Oriented object detection (OOD) in remote sensing images (RSIs) suffers from insufficient feature representation caused by arbitrary rotation angles and small spatial resolutions. Existing spatial–frequency fusion paradigms merely implement single-granularity feature interaction, either global image-level frequency compensation or local instance-level feature refinement, and [...] Read more.

Oriented object detection (OOD) in remote sensing images (RSIs) suffers from insufficient feature representation caused by arbitrary rotation angles and small spatial resolutions. Existing spatial–frequency fusion paradigms merely implement single-granularity feature interaction, either global image-level frequency compensation or local instance-level feature refinement, and fail to simultaneously capture global scene semantic consistency and local object fine-grained discriminability. In this paper, we propose a unified dual-level spatial–frequency collaborative detector (DSCDet) for remote sensing OOD tasks. Different from previous decoupled designs, the proposed DSCDet constructs a complete spatial–frequency collaborative fusion paradigm that shares a generic wavelet-based frequency extraction mechanism and cross-feature fusion module, which is adaptively deployed at both image-level and instance-level granularities. Specifically, our method introduces Haar wavelet transform to extract multi-scale frequency mutation features. On this basis, a generic cross-domain attention fusion (GCDAF) is constructed with granularity-dependent positional encoding constraints. The core difference between dual granularity fusion lies in geometric positional encoding, where image-level fusion adopts global scene positional embedding to maintain overall semantic stability, and instance-level fusion leverages local pairwise instance positional embedding to optimize fine-grained target feature interaction. The unified dual-level fusion architecture comprehensively integrates global semantic integrity and local target specificity, forming a robust and universal spatial–frequency feature representation system. Extensive experiments on three public remote sensing datasets, including DOTA-v1.0, DOTA-v1.5 and DIOR-R, demonstrate that the proposed DSCDet achieves competitive and superior performance against state-of-the-art OOD detectors. Full article

(This article belongs to the Section Remote Sensing Image Processing)

► Show Figures

Figure 1

25 pages, 59568 KB

Open AccessArticle

Mitigating Class Imbalance and False-Negative Supervision in Remote Sensing Semantic Segmentation Using Object-Centric Patch Sampling

by Yogesh Regmi, Sandeep Gautam, Gaurav Parajuli, Abinash Silwal, Roshan Bhandari and Tri Dev Acharya

Remote Sens. 2026, 18(16), 2844; https://doi.org/10.3390/rs18162844 - 21 Aug 2026

Viewed by 594

Abstract

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 [...] Read more.

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

(This article belongs to the Special Issue Remote Sensing Measurements of Land Use and Land Cover)

► Show Figures

Figure 1

34 pages, 2181 KB

Open AccessArticle

Weakly Supervised Remote Sensing Segmentation via Decoupled Cross-Modal Distillation and Semantic-Guided Refinement

by Jing Li, Yulin Cao, Xiantao Jiang, Dong Zhao and Dan Zhang

Remote Sens. 2026, 18(16), 2843; https://doi.org/10.3390/rs18162843 - 21 Aug 2026

Viewed by 409

Abstract

Pixel-level annotation of remote sensing imagery is costly, motivating weakly supervised semantic segmentation (WSSS) using only image-level labels. However, class activation maps (CAMs) often highlight only discriminative sub-regions and fail to separate adjacent land-cover regions, particularly in remote sensing scenes characterized by densely [...] Read more.

Pixel-level annotation of remote sensing imagery is costly, motivating weakly supervised semantic segmentation (WSSS) using only image-level labels. However, class activation maps (CAMs) often highlight only discriminative sub-regions and fail to separate adjacent land-cover regions, particularly in remote sensing scenes characterized by densely co-occurring land-cover classes and substantial variations in object scale. To address these limitations, we propose a three-stage framework that integrates complementary priors from Contrastive Language–Image Pre-training (CLIP), Self-Distillation with No Labels version 2 (DINOv2), and the Segment Anything Model (SAM). First, a lightweight CLIP adapter aligns vision–language priors with remote sensing imagery, while sigmoid-based multi-label decoupled distillation replaces class-competitive distillation with independent class-wise supervision, producing more complete CAMs. Second, DINOv2-guided feature clustering decomposes large merged regions before SAM prompt generation, while Spatial–Semantic Constraints are used to construct confidence-guided point-and-box prompts and reject excessively expanded or semantically inconsistent masks, thereby generating reliable pseudo-labels. Finally, a compact segmentation network is initialized with the weights learned in Stage 1 and retrained using the refined pseudo-labels generated in Stage 2, eliminating the need for foundation models during inference. Experiments on the Potsdam, LoveDA, and DeepGlobe datasets show that the proposed method achieves mean intersection over union (mIoU) scores of 53.16%, 52.66%, and 62.98%, respectively, outperforming state-of-the-art WSSS baselines by 6.55, 1.16, and 1.27 percentage points, respectively. These results demonstrate the effectiveness and generalizability of the proposed framework across diverse remote sensing scenarios under image-level supervision. Full article

(This article belongs to the Special Issue Learning-Based Remote Sensing and Earth Observation Intelligence: Representation, Fusion, Interpretation, and Generation)

► Show Figures

Figure 1

24 pages, 8958 KB

Open AccessArticle

HRRP Reconstruction Method for Coded Interrupted Sampling Radar Echoes Based on Multi-Frame Sequential Priors

by Ziai Zhang, Qihua Wu, Xiaobin Liu, Zhaoyu Gu, Shunping Xiao and Feng Zhao

Remote Sens. 2026, 18(16), 2842; https://doi.org/10.3390/rs18162842 - 21 Aug 2026

Viewed by 237

Abstract

High-resolution range profile (HRRP) reconstruction is essential for extracting range-direction scattering characteristics in wideband radar remote sensing, particularly in synthetic aperture radar (SAR) and inverse synthetic aperture radar (ISAR) imaging. Coded interrupted sampling (CIS) can improve radar low probability of intercept (LPI) performance [...] Read more.

High-resolution range profile (HRRP) reconstruction is essential for extracting range-direction scattering characteristics in wideband radar remote sensing, particularly in synthetic aperture radar (SAR) and inverse synthetic aperture radar (ISAR) imaging. Coded interrupted sampling (CIS) can improve radar low probability of intercept (LPI) performance by controlling signal transmission with a binary sequence. However, the reduced number of valid echo samples may degrade HRRP reconstruction, especially under low-duty-ratio and low signal-to-noise ratio (SNR) conditions. Conventional orthogonal matching pursuit (OMP) processes each frame independently and ignores the inter-frame continuity of scattering-center positions, which may lead to false selections and missed detections. To address this problem, this paper proposes a candidate-interval-assisted orthogonal matching pursuit (CI-OMP) algorithm based on multi-frame sequential priors. Stable scattering-center positions are extracted from historical reconstruction results and expanded into candidate intervals to guide atom matching in the current frame. Simulation results show that CI-OMP outperforms standard OMP in terms of normalized mean squared error (NMSE), tolerant support recovery rate (Tol-SRR), and peak-to-sidelobe ratio (PSLR). At a duty ratio of 0.20, CI-OMP reduces the NMSE by 1.71 dB and improves the PSLR by 7.56 dB compared with OMP. In addition, the candidate-interval strategy reduces the atom-search range by approximately 54–75% under different duty ratios and by approximately 50–83% under different SNRs, demonstrating improved search efficiency. These results demonstrate that CI-OMP improves the accuracy, robustness, and search efficiency of HRRP reconstruction for CIS radar echoes, particularly under low-duty-ratio and low-to-medium-SNR conditions. Full article

(This article belongs to the Special Issue Advances in Imaging Radar Signal Processing, Target Feature Extraction and Recognition)

► Show Figures

Figure 1

28 pages, 71265 KB

Open AccessArticle

Sharing Cultural Values Through 3D Point-Cloud-Based Documentation of Transylvanian Heritage

by Alina Elena Voinea, Calin Neamtu and Virgil Pop

Remote Sens. 2026, 18(16), 2841; https://doi.org/10.3390/rs18162841 - 21 Aug 2026

Viewed by 558

Abstract

This paper presents a pilot educational workflow that couples 3D remote sensing with heritage-driven pedagogy by engaging architecture master’s students in the documentation and digital archiving of Transylvanian cultural sites. Using terrestrial and mobile 3D scanning, students documented multiple typologies—wooden churches (Târgușor, Tioltiur), [...] Read more.

This paper presents a pilot educational workflow that couples 3D remote sensing with heritage-driven pedagogy by engaging architecture master’s students in the documentation and digital archiving of Transylvanian cultural sites. Using terrestrial and mobile 3D scanning, students documented multiple typologies—wooden churches (Târgușor, Tioltiur), historical ensembles (Mociu, Coplean), industrial sites (1 Mai–Luduș, Vânătorilor–Luduș), and an urban street segment (Potaissa)—to generate dense point clouds that served as the basis for geometric reconstruction, semantic interpretation, and condition assessment. The study describes how the characteristics of different construction systems (timber, brick, stone, mixed structures) relate to point-cloud quality, survey coverage, and subsequent CAD/BIM drafting, with attention to the qualitative reading of minor deformations in wooden churches and of degradation patterns in masonry and industrial buildings. We also consider how artefacts in the data (noise, occlusions, registration errors) affect scene understanding and the interpretation of derived observations relevant to condition assessment and, prospectively, to monitoring. For the Tioltiur dual-sensor case, the TLS and SLAM datasets were compared through an internal CloudCompare registration check (final RMS 0.1121 on 50,000 points, fixed scale 1.0 and theoretical overlap 100%), surface-density displays (r = 0.005 for the Z+F dataset and for the GeoSLAM dataset), fitted-wall-plane readings (dip values around 89 deg. and 85 deg.) and a longitudinal section documenting roof/vault deformation. Beyond technical performance, the paper examines the self-reported formative impact on students’ digital skills and their understanding of cultural values, arguing that participation in 3D data acquisition, processing, and interpretation positions them as co-creators of a living digital archive. Pre- and post-workshop questionnaires (n = 13 each) are analysed descriptively—counts, percentages and medians with interquartile ranges—because the two instruments are unmatched and carry no shared identifier, so no paired test is applied; post-workshop self-ratings of technical competence, heritage understanding, archival awareness and collaboration were consistently high (medians 4–5), with uneven access to VR the main gap. By connecting point-cloud-based documentation workflows with heritage education, the project outlines a transferable, monitoring-ready baseline model in which 3D remote sensing supports both careful documentation and the transmission of regional identity and cultural meaning in architectural training. As an exploratory pilot with a small, self-reported sample, the study reports descriptive and qualitative findings rather than validated metric or statistical results. Full article

(This article belongs to the Special Issue 3D Computer Vision for Remote Sensing: Point Clouds, Reconstruction and Scene Understanding)

► Show Figures

Figure 1

17 pages, 9346 KB

Open AccessArticle

Tracking Total Precipitable Water Vapor: A Multi-Instrument Comparative Analysis

by Rocio D. Rossi, Johan R. Villanueva Medina, Ricardo K. Sakai, Ujjawal Shah, Nakul N. Karle, Adrian Flores and Xiaowen Li

Remote Sens. 2026, 18(16), 2840; https://doi.org/10.3390/rs18162840 - 21 Aug 2026

Viewed by 368

Abstract

Atmospheric water vapor is a major driver of Earth’s climate, yet despite its vital role in driving extreme weather and informing Numerical Weather Prediction (NWP) models, precise quantification of Precipitable Water Vapor (PWV) remains a challenge due to its high spatial and temporal [...] Read more.

Atmospheric water vapor is a major driver of Earth’s climate, yet despite its vital role in driving extreme weather and informing Numerical Weather Prediction (NWP) models, precise quantification of Precipitable Water Vapor (PWV) remains a challenge due to its high spatial and temporal variability. To provide an upgraded evaluation reflecting the most recent data and next-generation instrumentation, this study evaluates the accuracy, relative to radiosonde measurements, in calculating PWV values across six different instruments: Microwave Radiometer (MWR), NOAA-21, TROPOMI, Pandora spectrometer, AERONET sun-photometer, and GNSS/GPS against the bias-corrected Vaisala RS-92 and RS-41 radiosondes over Beltsville, Maryland, utilizing an updated 2024–2025 database. As a certified GRUAN site, HUBC adheres to strict international observation standards designed specifically to provide reference-quality data and comprehensive corrections for systematic errors. This rigorous framework justifies their application as the definitive ‘referent truth’ benchmark for remote sensing validation. RMSE and bias were the primary metrics used to assess relative accuracy. GPS measurements provided the highest level of relative accuracy, yielding the lowest RMSE (1.50 mm) and a near-unity linear fit (y = 0.98x). NOAA-21 and TROPOMI exhibit higher random noise when compared to ground-based instrumentation, yet both obtain high relative accuracy retrievals with negligible biases. AERONET and Pandora also showed strong performance with low RMSEs and R² values of 0.986 and 0.988, respectively, while slightly underestimating PWV. In contrast, the Radiometer performed with the lowest relative accuracy, characterized by the highest RMSE (6.19 mm) and a significant negative bias (−5.55 mm). Although all instruments maintained high correlation coefficients (R² ≥ 0.904), these results indicate that satellite and ground-based remote sensing provide reliable PWV retrievals, while GPS presents the most robust benchmark for high-accuracy PWV retrievals relative to radiosonde measurements. The findings also underscore the need for instrument-specific calibration constants to better align remote sensing retrievals with in situ observations. Full article

(This article belongs to the Section Atmospheric Remote Sensing)

► Show Figures

Figure 1

31 pages, 3720 KB

Open AccessArticle

DA-GDNet: A Data-Augmented Gather-and-Distribute Network for Robust SAR Target Detection

by Feihong Zhao, Yanfeng Li, Wenqian Wu, Houjin Chen and Yujing Shang

Remote Sens. 2026, 18(16), 2839; https://doi.org/10.3390/rs18162839 - 21 Aug 2026

Viewed by 267

Abstract

Synthetic Aperture Radar (SAR) possesses the capacity for all-weather imaging and is widely applied in target detection. However, robust SAR target detection remains challenging due to the limited availability of task-relevant labeled samples that jointly cover target categories, depression angles, and complex target–background [...] Read more.

Synthetic Aperture Radar (SAR) possesses the capacity for all-weather imaging and is widely applied in target detection. However, robust SAR target detection remains challenging due to the limited availability of task-relevant labeled samples that jointly cover target categories, depression angles, and complex target–background contexts. In this paper, we propose a Data-Augmented Gather-and-Distribute Network (DA-GDNet) for SAR image target detection. By jointly optimizing at both the data and architectural levels, the proposed approach enhances the model’s capacity for target detection in complex backgrounds. Specifically, we design a SAR image data augmentation strategy that integrates three-dimensional modeling with deep learning. Meanwhile, we incorporate a Gather–Distribute (GD) mechanism and a Spatial Feature Enhancement Module (SFEM) to achieve efficient multi-scale feature fusion and enhance the saliency of target regions. Experimental results on the MSTAR dataset and ATRNet-STAR dataset demonstrate that DA-GDNet not only improves detection accuracy and robustness, but also significantly strengthens the model’s adaptability to variations in depression angles and complex backgrounds. Full article

(This article belongs to the Section AI Remote Sensing)

► Show Figures

Figure 1

33 pages, 2116 KB

Open AccessArticle

Hyper-VMIL: Topology-Aware Variational Hypergraph Multiple-Instance Learning for Weakly Supervised Hyperspectral Target Detection

by Haoran Hu, Weiyi Hu, Chengkang Duan and Zhao Yang

Remote Sens. 2026, 18(16), 2838; https://doi.org/10.3390/rs18162838 - 21 Aug 2026

Viewed by 426

Abstract

Region-level weakly supervised hyperspectral target detection (HTD) using multiple-instance learning (MIL) reduces annotation costs but encounters challenges such as bag label ambiguity, boundary over-smoothing, and test-time computational latency. To address these issues, we propose Hyper-VMIL, a spatial–spectral topology-regularized variational hypergraph network. Hyper-VMIL formulates [...] Read more.

Region-level weakly supervised hyperspectral target detection (HTD) using multiple-instance learning (MIL) reduces annotation costs but encounters challenges such as bag label ambiguity, boundary over-smoothing, and test-time computational latency. To address these issues, we propose Hyper-VMIL, a spatial–spectral topology-regularized variational hypergraph network. Hyper-VMIL formulates latent target localization as variational inference over dual-path hypergraphs: a boundary-aware spatial hypergraph modeling geometric patch continuity and a dynamic spectral-manifold hypergraph capturing non-local material similarity. Node-adaptive gating dynamically balances spatial and spectral evidence to mitigate over-smoothing near target boundaries. Furthermore, a confidence-aware continuous posterior refinement (CTPR) mechanism reduces the confirmation bias associated with conventional hard pseudo-label binarization. Finally, a teacher–student distillation strategy transfers contextual topology into a lightweight single-spectrum student detector. Benchmark experiments on simulated ASTER and airborne MUUFL Gulfport and Avon datasets show that Hyper-VMIL achieves competitive performance against 15 baseline methods. Notably, Hyper-VMIL supports dual inference modes: Context Mode provides improved detection accuracy (+4.6% average NAUC over VMIL-ECM on MUUFL), while Pixel Mode enables single-spectrum inference (

1.25 μ s

single-instance latency and an amortized streaming throughput of

0.015 μ s

per pixel) suitable for onboard real-time deployment. Full article

(This article belongs to the Special Issue Recent Advances in Object Detection with Hyperspectral Remote Sensing Data)

► Show Figures

Figure 1

25 pages, 26810 KB

Open AccessArticle

Development and Verification of an Automatic Tower-Based SIF Observation System Based on Narrow Field-of-View Scanning and DOAS Atmospheric Correction

by Chenyu Hu, Pinhua Xie, Zhaokun Hu, Haoxuan Feng and Ang Li

Remote Sens. 2026, 18(16), 2837; https://doi.org/10.3390/rs18162837 - 21 Aug 2026

Viewed by 292

Abstract

Sun-induced chlorophyll fluorescence (SIF) is an effective proxy for vegetation photosynthesis, but tower-based retrieval suffers from atmospheric path interference under humid and variable conditions. We present a DOAS-based SIF retrieval algorithm that operates in Fraunhofer lines (680–686 nm, 745–758 nm) and water vapour-sensitive [...] Read more.

Sun-induced chlorophyll fluorescence (SIF) is an effective proxy for vegetation photosynthesis, but tower-based retrieval suffers from atmospheric path interference under humid and variable conditions. We present a DOAS-based SIF retrieval algorithm that operates in Fraunhofer lines (680–686 nm, 745–758 nm) and water vapour-sensitive bands (717–727 nm). It constructs an adaptive reference spectrum from SCOPE simulations and PCA and incorporates H₂O absorption cross-sections into the fitting process for active atmospheric correction. The algorithm is implemented in a dedicated tower-based system integrating a 1° scanning gimbal with a high-resolution spectrometer. Validation with simulated and field data demonstrates the following: (1) the algorithm retrieves SIF with high fidelity (correlation coefficients >0.9 across all windows); (2) it exhibits lower water-vapour sensitivity and greater cloudy-sky stability than FLD, 3FLD, and SFM, achieving the lowest coefficient of variation (CV = 0.356); (3) over a complete wheat–rice rotation, the retrieved SIF tracks crop growth and phenological stages. This work provides a reliable solution for automated, high-precision tower-based SIF observation under complex atmospheric conditions. Full article

(This article belongs to the Section Remote Sensing in Agriculture and Vegetation)

► Show Figures

Figure 1

29 pages, 13723 KB

Open AccessArticle

High-Resolution Mapping and Interpretation of Stable Urban Surface CO₂ Concentration Patterns Using CSF-Processed Mobile Observations and Multiscale Remote Sensing in Shenzhen, China

by Guoxu Li, Tianle Sun, Yonglin Zhang, Hao Zhang, Lingyun Yao, Jianwen Zhang, Shiguang Xu, Wanjuan Song, Zheng Niu and Li Wang

Remote Sens. 2026, 18(16), 2836; https://doi.org/10.3390/rs18162836 - 21 Aug 2026

Viewed by 348

Abstract

High-resolution mapping of urban surface CO₂ is essential for refined carbon monitoring, emission management, and low-carbon urban planning. Mobile monitoring provides dense street-level observations, but raw CO₂ measurements are often affected by transient traffic disturbances, vehicle idling, and localized plume events, [...] Read more.

High-resolution mapping of urban surface CO₂ is essential for refined carbon monitoring, emission management, and low-carbon urban planning. Mobile monitoring provides dense street-level observations, but raw CO₂ measurements are often affected by transient traffic disturbances, vehicle idling, and localized plume events, which limits their direct use as stable spatial mapping targets. This study developed an integrated framework for predicting, mapping, and interpreting stable surface CO₂ patterns in Shenzhen by combining vehicle mobile observations, CSF processing, multiscale remote sensing predictors, machine learning. A CSF-based lower-envelope filter was used to suppress short-duration positive peaks and extract a more stable CO₂ accumulation signal from mobile observations. Multiscale predictors representing transportation, urban activity, surface environment, and built form were constructed to characterize both local and surrounding urban contexts. Compared with raw CO₂, the CSF-processed target substantially improved prediction performance. The best validation R² across the candidate models increased from 0.59 to 0.90 in April and from 0.62 to 0.93 in November. The predicted maps identified persistent high-CO₂ areas in central and southwestern Shenzhen. SHAP results showed that transport networks and urban activity reinforced surface CO₂ accumulation, whereas vegetation and open-surface contexts weakened accumulation at broader spatial ranges. These findings provide an interpretable framework for high-resolution urban CO₂ mapping and refined low-carbon governance. Full article

(This article belongs to the Special Issue Satellite Remote Sensing of Quantifying Greenhouse Gases Emissions)

► Show Figures

Figure 1

27 pages, 17769 KB

Open AccessArticle

SFSMamba-DETR: Selective Feature Scanning with State Space Models and Dual-Scale Window Attention for Remote Sensing Object Detection

by Yuanli Cai, Junchao Zhao, Husheng Wu and Rui Ma

Remote Sens. 2026, 18(16), 2835; https://doi.org/10.3390/rs18162835 - 21 Aug 2026

Viewed by 376

Abstract

Object detection in remote sensing imagery remains challenging due to vast scale variations, complex backgrounds, and the prevalence of small, densely packed targets. Existing CNN-based detectors are limited by restricted receptive fields, while Transformer-based methods incur prohibitive computational overhead for high-resolution inputs. In [...] Read more.

Object detection in remote sensing imagery remains challenging due to vast scale variations, complex backgrounds, and the prevalence of small, densely packed targets. Existing CNN-based detectors are limited by restricted receptive fields, while Transformer-based methods incur prohibitive computational overhead for high-resolution inputs. In this paper, we propose SFSMamba-DETR, a detection framework that integrates state space models with Dual-Scale Window Attention for efficient and accurate remote sensing object detection. Specifically, we design a Selective Feature Scanning (SFS) module that uses the Mamba-based 2D Selective Scan mechanism to model long-range spatial dependencies with linear computational complexity. To capture both fine-grained local patterns and broader contextual cues simultaneously, we introduce a Dual-Scale Window Attention (DSWA) mechanism that operates at two complementary window scales with multi-kernel convolution bridging. These modules are orchestrated within a Cross-scale Feature Aggregation Module (CFAM) that performs hierarchical multi-scale fusion in a hybrid encoder. Extensive experiments on three primary benchmarks (MAR20, UCAS-AOD, and the Jilin-1 Satellite Aircraft Detection Dataset), together with supplementary results on DOTA and DIOR, demonstrate that SFSMamba-DETR achieves strong detection accuracy while maintaining competitive inference speed. Full article

(This article belongs to the Special Issue Advanced Applications of Artificial Intelligence in Remote Sensing Image Recognition (2nd Edition))

► Show Figures

Figure 1

33 pages, 25847 KB

Open AccessArticle

Integrating Kernel-Based Vegetation Indices and Ensemble Learning for Mangrove Canopy Height Mapping Using GEDI and Sentinel Data

by Peilin Lai, Yang Chen, Wenqian Chen, Lixia Ma, Weijie Chen, Dongyang Fu, Dazhao Liu and Kai Tian

Remote Sens. 2026, 18(16), 2834; https://doi.org/10.3390/rs18162834 - 21 Aug 2026

Viewed by 367

Abstract

Mangrove canopy height (MCH) is a fundamental structural variable for monitoring ecosystem health and quantifying carbon stocks. However, MCH retrieval from optical satellite imagery is often constrained by spectral saturation in dense stands and environmental noise in intertidal zones. This study investigates the [...] Read more.

Mangrove canopy height (MCH) is a fundamental structural variable for monitoring ecosystem health and quantifying carbon stocks. However, MCH retrieval from optical satellite imagery is often constrained by spectral saturation in dense stands and environmental noise in intertidal zones. This study investigates the utility of kernel-based spectral features (KVIs) as non-linear topological enhancements for MCH estimation by integrating GEDI spaceborne LiDAR with Sentinel-2 and Sentinel-1 data across three mangrove ecosystems along the South China coast. Utilizing four regression models under spatial cross-validation, we evaluated the performance of traditional indices, KVIs, and integrated feature sets against GEDI reference measurements. Results indicate that traditional optical indices exhibit limited linear sensitivity to MCH. Rather than serving as universal accuracy boosters, KVIs function as non-linear stabilizers by redistributing spectral values in Hilbert space, which effectively enhances feature representation in high biomass stands and mitigates background noise. Furthermore, model comparisons reveal that while traditional indices yield competitive baseline accuracy in specific architectures (e.g., 1D-CNN), kernel-based features provide nuanced advantages in spatial stability and ensemble dispersion reduction. Ultimately, this study demonstrates that kernel-based features enhance model robustness under rigorous cross-validation, providing a reliable structural foundation for large-scale ecological monitoring and regional carbon dynamics assessments in heterogeneous coastal environments. Full article

(This article belongs to the Special Issue Multi-Source Remote Sensing Approaches for Monitoring and Conserving Forest Biodiversity: Advances in Methodology and Applications)

► Show Figures

Figure 1

27 pages, 4364 KB

Open AccessArticle

Spatial–Spectral Decoupling-Enhanced Lightweight Network for Few-Shot Hyperspectral Anomaly Detection in Remote Sensing Imagery

by Hongwei Qu, Qing Guo and Jinlin Zou

Remote Sens. 2026, 18(16), 2833; https://doi.org/10.3390/rs18162833 - 20 Aug 2026

Viewed by 519

Abstract

Hyperspectral anomaly detection (HAD) identifies targets by spectral differences. However, large deep learning models overfit under the small-sample conditions typical of remote sensing, where anomalies are sparse and annotations costly. We propose a lightweight network on the GT-HAD transformer backbone for hyperspectral imagery. [...] Read more.

Hyperspectral anomaly detection (HAD) identifies targets by spectral differences. However, large deep learning models overfit under the small-sample conditions typical of remote sensing, where anomalies are sparse and annotations costly. We propose a lightweight network on the GT-HAD transformer backbone for hyperspectral imagery. The design includes: (1) capacity reduction via multi-layer perceptron (MLP) ratio and embedding dimension optimization; (2) a Decoupled Projection Gating Fusion (DPGF) module enforcing spatial–spectral decoupling via dual-path projection and complementary regularization; (3) analysis of capacity–generalization tradeoffs across six airborne hyperspectral datasets. Experiments reveal an inverted-U relationship between capacity and generalization. The Mini configuration (121 K parameters) achieves a 53% parameter reduction. It yields AUC improvements on three datasets, maintains negligible performance loss (<0.1%) on two datasets, and exhibits a measurable decline on only one dataset, compared with the 255 K baseline. Notably, the baseline attains higher accuracy with only 25% training data on Pavia (+3.69% area under the receiver operating characteristic curve, AUC). This dataset-specific observation provides empirical evidence that capacity–data mismatch can induce severe overfitting in deep HAD models. Under a 25% training ratio, DPGF shows observable performance trends on spectrally homogeneous scenes, and zero initialization ensures identical performance to that of the baseline at the initial training stage, eliminating insertion risk during module deployment. However, the limited diversity of sensors and scene types constrains the generalizability of the observed trends. These results demonstrate that spatial–spectral decoupling with lightweight design suppresses overfitting in few-shot HAD, guiding compact models for onboard satellite and unmanned aerial vehicle (UAV) applications. Full article

► Show Figures

Figure 1

18 pages, 16023 KB

Open AccessArticle

Multi-Source Geophysical Data Integration for Underwater Target Detection in Complex Seabed Environments: A Case Study of the Nan’ao I Shipwreck, China

by Yonghang Li, Jiale Chen, Yuanzhao Meng, Dashun Xiao, Hai Lin, Huiqiang Yao, Zepeng Huang, Haoyi Zhou and Shi Zhang

Remote Sens. 2026, 18(16), 2832; https://doi.org/10.3390/rs18162832 - 20 Aug 2026

Viewed by 493

Abstract

The search and discovery of underwater shipwreck sites represent the most arduous and critical phases of underwater archaeology. Wooden shipwrecks, in particular, are characterized by low acoustic impedance contrast and weak magnetic anomalies, coupled with their limited physical dimensions. Consequently, they predominantly exist [...] Read more.

The search and discovery of underwater shipwreck sites represent the most arduous and critical phases of underwater archaeology. Wooden shipwrecks, in particular, are characterized by low acoustic impedance contrast and weak magnetic anomalies, coupled with their limited physical dimensions. Consequently, they predominantly exist as shallow-buried, discontinuous small targets scattered within confined areas, making their detection exceptionally challenging. Furthermore, the complexity of the submarine environment—including rugged topography, turbid water columns, and strong currents—poses formidable obstacles to the effective detection of these archaeological remains. Single geophysical methods are often limited by insufficient imaging resolution, interpretation ambiguity, and geological noise, making precise localization and characterization difficult. Focusing on the Nan’ao I Ming Dynasty shipwreck located in waters approximately 24 m deep off the coast of Nan’ao, Guangdong Province, China, this study proposes and validates an “acoustic-magnetic” multi-source data integration detection method. This approach systematically integrates high-resolution multibeam echo sounding (MBES), side-scan sonar (SSS), sub-bottom profiling (SBP), and marine magnetic data to establish a comprehensive framework for identification and integration analysis. The results indicate that the MBES bathymetric data reveal a regular, elongated structure oriented north–south (approximately 34 m × 12 m), closely matching the main hull and deck configuration. The SSS imagery exhibited high backscatter intensity and parallel linear textures, effectively delineating the hard shipwreck structure and the associated rigid protective frame employed for in situ preservation. SBP data confirmed the semi-buried state of the shipwreck (burial depth of approximately 0.6 m). Spatial variations in sediment thickness around the site suggested ongoing modification by strong hydrodynamic processes. Marine magnetic surveys identified localized negative anomalies (−210 nT relative to the ambient magnetic field), contrasting sharply with the positive anomalies of the surrounding natural reefs, thereby indicating an artificial ferromagnetic source. The spatial registration and feature superposition of multi-source data facilitated the characterization of the shipwreck, demonstrating its potential to mitigate environmental interference and enhance detection reliability in this complex environment. Using the Nan’ao I shipwreck site as a case study, this study provides a detailed characterization of the site’s 3D morphology, burial state, and physical properties. The proposed methodology offers a practical and robust technical solution for underwater shipwreck archaeology in complex nearshore environments, providing significant implications for proactive discovery, efficient investigation, and protection of underwater cultural heritage (UCH). Full article

(This article belongs to the Special Issue Remote, Proximal Sensing and Geophysics for Cultural Heritage Knowledge and Conservation (Second Edition))

► Show Figures

Figure 1

29 pages, 14947 KB

Open AccessArticle

Multiscale Estimation of Mangrove Biomass in Fujian Province Using UAV as a Bridging Scale

by Shuwei Chen, Xi He, Yingbin Zhang, Xinhuang Zhang, Zhichao Cai and Riwen Lai

Remote Sens. 2026, 18(16), 2831; https://doi.org/10.3390/rs18162831 - 20 Aug 2026

Viewed by 378

Abstract

Mangroves are important coastal blue-carbon ecosystems, and accurate biomass estimation is essential for carbon stock assessment and ecological monitoring. To address the limitation of regional-scale biomass estimation caused by the scale mismatch between field plots and satellite pixels, this study selected the Zhangjiangkou [...] Read more.

Mangroves are important coastal blue-carbon ecosystems, and accurate biomass estimation is essential for carbon stock assessment and ecological monitoring. To address the limitation of regional-scale biomass estimation caused by the scale mismatch between field plots and satellite pixels, this study selected the Zhangjiangkou National Mangrove Nature Reserve in Fujian Province as the study area and the mangrove distribution region of Fujian Province as the extrapolation area. A multiscale biomass estimation framework integrating field plots, unmanned aerial vehicles (UAVs), and satellite remote sensing was established. The results showed that (1) the optimal UAV-scale models achieved R² values of 0.69 and 0.78 for aboveground biomass (AGB) and belowground biomass (BGB), respectively, with corresponding root mean square error (RMSE) values of 18.55 and 9.52 t·ha⁻¹ and normalized root mean square error (nRMSE) values of 0.14 and 0.17 demonstrating reliable predictive performance; (2) after introducing UAV-derived bridging labels, the R² of the AGB model increased from 0.24 to 0.64, while the RMSE decreased from 29.02 to 10.86 t·ha⁻¹. Similarly, the R² of the BGB model increased from 0.43 to 0.63, accompanied by a reduction in RMSE from 14.89 to 6.35 t·ha⁻¹, demonstrating a substantial improvement in satellite-scale biomass estimation accuracy; (3) the total AGB and BGB of mangroves in Fujian Province were estimated at 58,768.60 t and 24,575.14 t, respectively, with high-biomass areas mainly distributed along the coastal regions of Zhangzhou and Quanzhou. Unlike conventional field-to-satellite extrapolation approaches, the proposed framework introduces UAV-derived biomass maps as intermediate bridging labels for pixel-level supervised learning, thereby establishing an effective link between field measurements and satellite observations. This strategy effectively reduces the scale mismatch between field and satellite data, significantly improves satellite-scale biomass estimation accuracy, and provides a transferable and scalable framework for regional mangrove biomass mapping and blue-carbon assessment. Full article

► Show Figures

Figure 1

27 pages, 5444 KB

Open AccessArticle

Evaluating Seasonal Fidelity and Cross-Site Structural Discrimination of Sentinel-2 LAI Products in Karst Forests

by Magdalena Năpăruş-Aljančič, Alina L. Machidon, Urša Vilhar, Erika Kozamernik, Lado Kutnar, Janez Kermavnar, Žan Kafol, Nataša Ravbar and Tanja Pipan

Remote Sens. 2026, 18(16), 2830; https://doi.org/10.3390/rs18162830 - 20 Aug 2026

Viewed by 498

Abstract

Leaf area index (LAI) is widely used to characterize foliage amount and seasonal canopy development, but it captures only selected aspects of forest structure and can be difficult to retrieve reliably in heterogeneous, multilayered stands. This study evaluates Sentinel-2-based LAI information across eight [...] Read more.

Leaf area index (LAI) is widely used to characterize foliage amount and seasonal canopy development, but it captures only selected aspects of forest structure and can be difficult to retrieve reliably in heterogeneous, multilayered stands. This study evaluates Sentinel-2-based LAI information across eight sites in the Slovenian Classical Karst encompassing post-disturbance regeneration and established forest stands in dolines and relatively level inter-doline terrain. Field effective LAI measured during six periods in 2021 was compared with six Sentinel-2 spectral variables, LAI derived using the Sentinel Application Platform (SNAP), and the Copernicus Land Monitoring Service High-Resolution LAI product. The analysis explicitly distinguished two dimensions of retrieval performance that are often conflated: seasonal fidelity within sites and preservation of structural differences among sites. Most satellite-derived variables and LAI products captured the broad phenological progression from canopy development to senescence. However, strong temporal agreement within sites did not consistently translate into preservation of the ordering or magnitude of structural differences among sites. Several methods compressed the range of high effective LAI values at dense regeneration sites with substantial lower-layer vegetation. The study therefore provides a more informative framework for evaluating LAI products by identifying which component of variation drives apparent agreement. These findings indicate that Sentinel-2 can support phenological monitoring and broad screening of post-disturbance vegetation development. However, quantitative comparisons of canopy density or structural recovery across heterogeneous stands require consideration of canopy heterogeneity, potential spectral saturation, and plot-to-pixel support. The evaluation framework and the observed retrieval limitations are relevant beyond karst forests, particularly to post-disturbance stands, fragmented forests, open woodlands, and sites with dense understory or regeneration layers. Full article

(This article belongs to the Special Issue Perspectives of Remote Sensing to Assess Forest Structure and Diversity)

► Show Figures

Figure 1

25 pages, 20950 KB

Open AccessArticle

Long-Horizon Mining Subsidence Forecasting and Ecological Time-Lag Assessment Using Multi-Source Remote Sensing

by Lei Zhang, Lijun Duan and Shangmin Zhao

Remote Sens. 2026, 18(16), 2829; https://doi.org/10.3390/rs18162829 - 20 Aug 2026

Viewed by 283

Abstract

Surface deformation induced by underground coal mining is characterized by strong nonlinearity and spatial heterogeneity, which complicates early warning and ecological assessment. While InSAR-driven data assimilation models and remote sensing-based ecological indices are widely used for long-term monitoring, three fundamental limitations remain unresolved: [...] Read more.

Surface deformation induced by underground coal mining is characterized by strong nonlinearity and spatial heterogeneity, which complicates early warning and ecological assessment. While InSAR-driven data assimilation models and remote sensing-based ecological indices are widely used for long-term monitoring, three fundamental limitations remain unresolved: (1) severe spatial imbalance in deformation samples biases data-driven models toward mean-reverting predictions, (2) recursive multi-step forecasting accumulates errors, leading to instability in long-horizon extrapolation, and (3) in ecological monitoring, vegetation resilience further induces a multi-year observation lag, resulting in a “pseudo-stable” bias in optical indicators. To address these issues, this study proposes an unified framework integrating multi-step deformation prediction and ecological time-lag analysis. Taking the Datong Coalfield as the study area, we utilized 231 Sentinel-1A images from March 2017 to December 2024 for SBAS-InSAR deformation inversion. A spatial stratified sampling strategy is used to extract 5894 representative points. A 24-step backward and 15-step forward windows were reconstructed to systematically compare six predictive models. Simultaneously, the Remote Sensing Ecological Index (RSEI) derived from Landsat data is used for cross-lagged analysis. The results demonstrate that: (1) The maximum deformation rate reached −276.75 mm/year, with cumulative subsidence exceeding −2000 mm. (2) At 3-step short-term forecasting, all models proved robust, with LSTM performing best (RMSE = 5.78 mm). At 15-step extreme extrapolation, however, traditional recursive models diverged significantly (Kalman, RMSE = 45.70 mm), whereas N-BEATS maintained stability and effectively mitigated temporal error cascades with an RMSE of 17.98 mm. (3) The core collapse zone exhibited concurrent ecological degradation (Lag 0), while the marginal basin presented a hidden degradation period of one to two years. It provides reliable scientific support for precise tracking and proactive safety management in complex mining areas. Full article

(This article belongs to the Special Issue Recent Advances in InSAR for Deformation Monitoring and Earth Surface Process Analysis)

► Show Figures

Figure 1

26 pages, 9396 KB

Open AccessArticle

Multi-Scale Spatiotemporal Graph ODE Networks for Marine Chlorophyll-a Prediction

by Xiaoyu He, Yijing Zhang, Xin Huang and Suixiang Shi

Remote Sens. 2026, 18(16), 2828; https://doi.org/10.3390/rs18162828 - 20 Aug 2026

Viewed by 273

Abstract

Chlorophyll-a concentration is a key indicator reflecting the growth status of phytoplankton, and its accurate prediction is of great significance for assessing the degree of water eutrophication. Although existing approaches have achieved good performance, they generally pay insufficient attention to multi-scale spatial information [...] Read more.

Chlorophyll-a concentration is a key indicator reflecting the growth status of phytoplankton, and its accurate prediction is of great significance for assessing the degree of water eutrophication. Although existing approaches have achieved good performance, they generally pay insufficient attention to multi-scale spatial information and show limitations in characterizing the continuous spatiotemporal dynamics. To address these issues, this paper proposes a multi-scale spatiotemporal graph ODE network (MGODE) for ocean chlorophyll-a prediction. The MGODE adopts a dual-layer structure, simultaneously processing chlorophyll-a concentration data at both the region level and node level to capture multi-scale spatial features, and it enables effective interaction of cross-scale features through dynamic transmission coefficients and a gated fusion mechanism. Meanwhile, the MGODE employs a dual-ODE architecture at both the node and region levels, utilizing spatiotemporal ODE blocks to continuously and deeply capture features, thereby simulating the continuous spatiotemporal dynamic evolution of chlorophyll-a. Experiments on real-world datasets from the Bohai Sea and South China Sea show that the proposed MGODE model achieves higher prediction accuracy than several current state-of-the-art models. Compared with the best baseline, the MGODE achieves reductions of 2.78% in MAE and 1.07% in RMSE on the Bohai Sea dataset and reductions of 1.19% in MAE and 1.38% in RMSE on the South China Sea dataset. These results demonstrate the potential of the MGODE to support marine chlorophyll-a forecasting and marine ecological monitoring. Full article

(This article belongs to the Special Issue Advancing Ocean Observation, Analysis, and Forecasting Through AI-Powered Remote Sensing)

► Show Figures

Figure 1

28 pages, 12985 KB

Open AccessArticle

GCF-Net: Stage-Aligned Optical–Elevation Fusion for Aerial Remote Sensing Semantic Segmentation

by Yifan Yu, Song Deng, Yang Yang and Fan Min

Remote Sens. 2026, 18(16), 2827; https://doi.org/10.3390/rs18162827 - 20 Aug 2026

Viewed by 403

Abstract

Optical–elevation data fusion is widely used in aerial remote sensing semantic segmentation, as optical imagery provides rich spectral and textural information, while DSM or DEM data offer complementary elevation-related structural cues. However, effective fusion remains challenging because optical and elevation representations may exhibit [...] Read more.

Optical–elevation data fusion is widely used in aerial remote sensing semantic segmentation, as optical imagery provides rich spectral and textural information, while DSM or DEM data offer complementary elevation-related structural cues. However, effective fusion remains challenging because optical and elevation representations may exhibit cross-modal structural inconsistency, frequency–spatial response imbalance, and decoder-stage structural attenuation. To address these challenges, we propose GCF-Net, a stage-aligned optical–elevation fusion network that matches different cross-modal processing objectives to the evolving representation states of the encoder–decoder pipeline. A Structure-Guided Cross-Modal Correction Module first performs structure-conditioned correction of modality-specific features before fusion. A Frequency–Spatial Cross-Modal Fusion Module then constructs joint representations through bounded cross-modal magnitude conditioning, frequency-to-spatial reconstruction, and spatial recalibration. During decoding, a Geometry-Aware Cross-Scale Refinement Module reintroduces elevation-derived structural guidance into multiscale fused features. Experiments on ISPRS Vaihingen, ISPRS Potsdam, and MMHunan yield mIoU scores of 72.37%, 75.38%, and 52.23%, respectively, achieving the highest mIoU among the evaluated unimodal, multimodal, and SAM-based methods under the unified protocol. Ablation and replacement experiments verify the complementary roles of the three stage-specific components, while sensitivity, elevation perturbation, and complexity analyses indicate architectural flexibility, tolerance to moderate elevation degradation, and a balanced accuracy–efficiency trade-off. Full article

(This article belongs to the Section Remote Sensing Image Processing)

► Show Figures

Figure 1

27 pages, 2665 KB

Open AccessArticle

Midday Depression and Legacy Effect Disrupt SIF-GPP Coupling in Northern Peatlands During Combined Heat and Drought Stress

by Abdallah Yussuf Ali Abdelmajeed, M.Pilar Cendrero-Mateo, Shari Van Wittenberghe, Michal Antala, Mar Albert-Saiz, Marcin Stróżecki, Anshu Rastogi, Tommaso Julitta, Andreas Burkart, Dirk Schuettemeyer, Sheng Wang and Radosław Juszczak

Remote Sens. 2026, 18(16), 2826; https://doi.org/10.3390/rs18162826 - 20 Aug 2026

Viewed by 369

Abstract

Peatlands, critical global carbon sinks, are facing increasing threats from climate change-driven heatwaves and droughts. These threats can cause a midday depression in carbon uptake through photosynthetic inhibition. Using high-temporal-resolution solar-induced chlorophyll fluorescence (SIF; ~30 s) and chamber-based CO₂ flux measurements, we [...] Read more.

Peatlands, critical global carbon sinks, are facing increasing threats from climate change-driven heatwaves and droughts. These threats can cause a midday depression in carbon uptake through photosynthetic inhibition. Using high-temporal-resolution solar-induced chlorophyll fluorescence (SIF; ~30 s) and chamber-based CO₂ flux measurements, we investigated the coupling between SIF and gross primary production (GPP) during extreme events (air temperature > 25 °C and vapour pressure deficit > 15 hPa) in a northern peatland. Our results show that SIF tracks GPP closely under non-stress conditions (daily R² = 0.86–0.96). However, during combined heat and drought stress, midday correlations collapsed (Case A: R² = 0.04 on 27 June; Case B: R² = 0.15 and 0.01 on 29 and 30 June, respectively), indicating severe decoupling. Importantly, we discovered legacy effects from multi-day heat exposure: on 26 June, vegetation with prior cumulative stress (Case A) showed weak morning coupling (R² = 0.07), while vegetation without prior stress history (Case B) maintained strong coupling (R² = 0.93). This suggests that cumulative stress alters baseline physiology and can exacerbate midday mismatches; therefore, not just current condition controls photosynthetic regulation. These findings highlight limitations of SIF-based GPP estimation at sub-daily timescales during stress, particularly in heterogeneous peatland systems where canopy composition and physiological responses could vary among plant functional types. Full article

(This article belongs to the Section Remote Sensing in Agriculture and Vegetation)

► Show Figures

Figure 1

21 pages, 2427 KB

Open AccessArticle

Enhancing the Reliability, Performance, and Practical Value of Deep Learning Landslide Mapping via Pixel-Level Uncertainty Analysis

by Shan Dong, Hui Li, Cheng Zhong and Junjie Pan

Remote Sens. 2026, 18(16), 2825; https://doi.org/10.3390/rs18162825 - 20 Aug 2026

Viewed by 367

Abstract

Pixel-level uncertainty assessment is commonly overlooked in current deep learning-based landslide detection, which greatly limits the reliability, performance, and practical value of the results. To fill this gap, this study employs Monte Carlo Dropout to systematically quantify pixel-level uncertainty. Subsequently, it interprets how [...] Read more.

Pixel-level uncertainty assessment is commonly overlooked in current deep learning-based landslide detection, which greatly limits the reliability, performance, and practical value of the results. To fill this gap, this study employs Monte Carlo Dropout to systematically quantify pixel-level uncertainty. Subsequently, it interprets how external factors, such as terrain, vegetation, clouds, and shadows, interfere with model predictions. Taking SegFormer as an example, this study further investigates how weight allocation across different scales influences the uncertainty. Finally, we introduce an uncertainty ranking-based FP rejection strategy coupled with an FN priority capture strategy to improve the efficiency of mapping results inspection, thus improving landslide identification performance. The results indicate that by removing only the top 10% of pixels with the highest uncertainty, the mIoU increases by at least 7%. This study greatly enhanced the reliability, performance, and practical value of remote sensing-based landslide mapping from a new perspective. Full article

(This article belongs to the Special Issue Geohazard Mapping, Monitoring and Prediction with Advanced Remote Sensing Techniques)

► Show Figures

Figure 1

27 pages, 36264 KB

Open AccessArticle

Multiple Vegetation Indicators Reveal Contrasting Post-Drought Recovery Time in the Yangtze River Basin Following the 2022 Extreme Drought

by Qingqing Ma, Lajiao Chen, Jiepeng Li and Peng Liu

Remote Sens. 2026, 18(16), 2824; https://doi.org/10.3390/rs18162824 - 20 Aug 2026

Viewed by 302

Abstract

Extreme drought events have become increasingly frequent under ongoing climatic change, thereby constraining vegetation growth and altering ecosystem processes. Vegetation recovery time following drought plays a crucial role in ecosystem stability, and extensive studies have been conducted to quantify vegetation recovery. However, most [...] Read more.

Extreme drought events have become increasingly frequent under ongoing climatic change, thereby constraining vegetation growth and altering ecosystem processes. Vegetation recovery time following drought plays a crucial role in ecosystem stability, and extensive studies have been conducted to quantify vegetation recovery. However, most studies estimate vegetation recovery time using a single vegetation index, which does not adequately reflect how vegetation responds to drought conditions, since different vegetation indicators reflect different facets of vegetation dynamics. In this study, multiple vegetation indicators, including Normalized Difference Vegetation Index (NDVI), Enhanced Vegetation Index (EVI), Leaf Area Index (LAI), Gross Primary Productivity (GPP), and Solar-Induced Chlorophyll Fluorescence (SIF), were applied to investigate post-drought vegetation recovery in the Yangtze River Basin (YRB). The findings reveal that: (1) most vegetation recovered within four months after drought, with one-month recovery being the most prevalent, followed by four-month recovery; (2) the average recovery times derived from EVI, LAI, NDVI, GPP, and SIF were 2.23, 1.62, 2.45, 2.79, and 2.00 months respectively; (3) forests exhibited the fastest recovery rates, whereas shrublands recovered the slowest. This study assesses post-drought vegetation status via the recovery duration, offers theoretical basis for water resource allocation optimization, and provides important reference for coping with future ecological risks. Full article

(This article belongs to the Section Environmental Remote Sensing)

► Show Figures

Figure 1

30 pages, 81014 KB

Open AccessArticle

AERO: Arbitrary-Scale Equivariant Resolution Operator for Remote Sensing Image Super-Resolution

by Rui Qin, Ying Shi and Yuhan Liu

Remote Sens. 2026, 18(16), 2823; https://doi.org/10.3390/rs18162823 - 20 Aug 2026

Viewed by 342

Abstract

Remote sensing image super-resolution aims to reconstruct high-resolution images from low-resolution observations and is important for image interpretation. Existing fixed-scale methods achieve good performance at predefined integer scales, but their dedicated upsampling modules limit their application to arbitrary-scale scenarios such as interactive GIS [...] Read more.

Remote sensing image super-resolution aims to reconstruct high-resolution images from low-resolution observations and is important for image interpretation. Existing fixed-scale methods achieve good performance at predefined integer scales, but their dedicated upsampling modules limit their application to arbitrary-scale scenarios such as interactive GIS and multi-source image fusion. Continuous implicit methods provide scale flexibility but often exhibit spectral bias, resulting in over-smoothed textures and blurred object boundaries. To overcome these limitations, we propose an Arbitrary-scale Equivariant Resolution Operator (AERO) for remote sensing image super-resolution. AERO consists of three components. The Omnidirectional Feature Extractor enhances feature representation under orientation variations. The Wavelet–Arnold Residual Group models low- and high-frequency information in the wavelet domain to preserve textures and geographic boundaries. The Local Implicit Terrain Operator employs relative sub-pixel coordinates for continuous arbitrary-scale reconstruction. Experiments on AID, NWPU-RESISC45, UCMerced, and WHU-RS19 demonstrate that AERO achieves the best performance in the ×4 fixed-scale task. On WHU-RS19, AERO reaches a PSNR of 31.02 dB, exceeding FMSR by 0.64 dB. In rotational robustness experiments, the maximum PSNR fluctuation is reduced from 0.0181 dB to 0.0010 dB. The results show that AERO provides a practical approach for arbitrary-scale remote sensing image super-resolution. Full article

(This article belongs to the Special Issue AI-Driven Remote Sensing Image Restoration and Generation)

► Show Figures

Figure 1

26 pages, 45260 KB

Open AccessArticle

Asynchronous Responses of Ecosystem Carbon Gain and Groundwater Storage Under Ecological Restoration in the Loess Plateau

by Yifei Ma, Qiaoli Wu, Shaoyuan Chen, Jinling Song and Jie Jiang

Remote Sens. 2026, 18(16), 2822; https://doi.org/10.3390/rs18162822 - 20 Aug 2026

Viewed by 330

Abstract

Since the implementation of the Grain-for-Green Program (GGP), vegetation across the Loess Plateau (LP) has substantially recovered. However, whether the associated increase in ecosystem carbon gain was accompanied by a proportional increase in water consumption and whether groundwater storage changed synchronously remain unclear. [...] Read more.

Since the implementation of the Grain-for-Green Program (GGP), vegetation across the Loess Plateau (LP) has substantially recovered. However, whether the associated increase in ecosystem carbon gain was accompanied by a proportional increase in water consumption and whether groundwater storage changed synchronously remain unclear. This study integrated multi-source remote sensing products, GLDAS-Noah land-surface assimilation data, GRACE/GRACE-FO satellite gravimetry, irrigation water-use data, provincial water-use statistics, and coal-resource information to examine long-term changes in gross primary productivity (GPP), evapotranspiration (ET), water-use efficiency (WUE), soil moisture (SM), and groundwater storage anomaly (GWSA) during 2002–2023. GPP increased significantly by 10.67 g C m⁻² yr⁻¹ (

p < 0.01

), whereas ET increased more modestly by 1.97 mm yr⁻¹ (

p < 0.05

). The relative growth rate of GPP (1.66%) was approximately 3.5 times that of ET (0.47%), and WUE increased by 0.018 g C m⁻² mm⁻¹ yr⁻¹ (

p < 0.01

). In the XGBoost–SHAP models for 2004–2019, LAI showed the strongest model-based association with GPP and WUE, whereas ET was associated more broadly with LAI, air temperature, and precipitation. SM declined during 2002–2015 but showed an increasing tendency during 2016–2023, particularly in the middle and deep layers. The long-term GWSA slopes derived from CSR and JPL were −8.707 and −9.505 mm yr⁻¹, respectively, and the averaged CSR–JPL GWSA series showed a Sen’s slope of −9.131 mm yr⁻¹. GWSA declined during 2002–2020 and showed only a short-term, nonsignificant increase during 2020–2023 (4.110 mm yr⁻¹,

p > 0.05

). These contrasting trajectories indicate that increases in surface carbon uptake and improvements in soil-water conditions were not accompanied by synchronous regional groundwater recovery. Overall, the ecological-restoration period was accompanied by increased carbon gain and WUE without a proportional increase in regional ET, while groundwater storage followed a distinct trajectory. These findings provide regional-scale evidence and a quantitative basis for coordinating sustainable water-resource management with ecological-restoration optimization on the LP. Full article

► Show Figures

Figure 1

33 pages, 6003 KB

Open AccessArticle

Unsupervised Gaussian-Noise-Robust Remote Sensing Change Detection via FRFCM-IRM Change Intensity Modeling and SEEDSAM-Constrained HCRF

by Lei Fan, Jiaxin Song, Yikun Li, Yuxi Hu and Yingang Ren

Remote Sens. 2026, 18(16), 2821; https://doi.org/10.3390/rs18162821 - 20 Aug 2026

Viewed by 270

Abstract

Remote sensing change detection technology is widely used in land-use monitoring, urban planning, and disaster assessment. However, during imaging and transmission, bi-temporal remote sensing images are vulnerable to Gaussian noise, which makes it difficult for change detection algorithms to distinguish truly changed areas [...] Read more.

Remote sensing change detection technology is widely used in land-use monitoring, urban planning, and disaster assessment. However, during imaging and transmission, bi-temporal remote sensing images are vulnerable to Gaussian noise, which makes it difficult for change detection algorithms to distinguish truly changed areas from noise-affected regions. To address this issue, this study proposes an unsupervised Gaussian-noise-robust change detection algorithm, termed FRIH-SEEDSAM. The proposed method first applies the Fast and Robust Fuzzy C-Means (FRFCM) algorithm to perform noise-resistant fuzzy clustering on bi-temporal remote sensing images. To establish reliable correspondences between the clustering results, the Integrated Region Matching (IRM) algorithm is introduced to construct weighted matching relationships while reducing the influence of abnormal memberships. The change intensity of spatially corresponding pixels is then calculated to generate a more stable change intensity map. Subsequently, the change intensity map is input into the Hybrid Conditional Random Field (HCRF) to infer pixel-level change labels, where the object potential function is constructed from the segmentation results of the Superpixels Extracted via Energy-Driven Sampling (SEEDS)-guided Segment Anything Model (SEEDSAM), which uses the centroids of the SEEDS superpixel regions as point prompts for the SAM, thereby enhancing change-label consistency within the same changed object region. The experimental results show that the FRIH-SEEDSAM algorithm maintains stable change detection performance across different datasets and under varying Gaussian noise levels. It outperforms the comparison algorithms in terms of several accuracy evaluation indicators, including Kappa and F1. Furthermore, even when the Gaussian noise variance increases to 0.05, Kappa remains at 0.8 or above on multiple dataset images. Full article

► Show Figures

Figure 1

22 pages, 2602 KB

Open AccessArticle

A Balanced Spectral–Spatial Cross-Fusion Network for Hyperspectral Anomaly Detection

by Yuquan Gan, Mengjiao Wang, Lei Zhang, Weidong Zhang, Tao Yu and Hongwei Wang

Remote Sens. 2026, 18(16), 2820; https://doi.org/10.3390/rs18162820 - 20 Aug 2026

Viewed by 450

Abstract

Hyperspectral anomaly detection aims to find abnormal targets without prior information. However, the high dimensionality of hyperspectral data, complicated spatial structures, and varying object scales make it challenging to jointly utilize spectral and spatial information. Therefore, anomalies may be confused with background regions. [...] Read more.

Hyperspectral anomaly detection aims to find abnormal targets without prior information. However, the high dimensionality of hyperspectral data, complicated spatial structures, and varying object scales make it challenging to jointly utilize spectral and spatial information. Therefore, anomalies may be confused with background regions. A Balanced Spectral–Spatial Cross-Fusion Network (BSCF-Net) is proposed for hyperspectral anomaly detection. The network uses a multi-branch encoder, where spatial branches capture features with different receptive fields and spectral branches extract spectral patterns through one-dimensional convolutions and channel attention. The Bidirectional Spectral–Spatial Cross-Attention (BSCA) mechanism enables information exchange between spectral and spatial features. The Multi-Scale Gated Refiner (MSGR) module is used to refine the fused features. With an autoencoder reconstruction framework, BSCF-Net identifies anomalies according to reconstruction errors and reduces background interference. Experimental results on five public hyperspectral datasets demonstrate the effectiveness of BSCF-Net, achieving competitive AUC performance under diverse background conditions. Full article

(This article belongs to the Section Earth Observation Data)

► Show Figures

Figure 1

25 pages, 7379 KB

Open AccessArticle

Improving Land Surface Emissivity for Better Simulation of Microwave Radiances over Northern Latitudes

by Swapan Mallick, Stéphanie Guedj and Magnus Lindskog

Remote Sens. 2026, 18(16), 2819; https://doi.org/10.3390/rs18162819 - 20 Aug 2026

Viewed by 441

Abstract

The utilisation of microwave radiances is crucial for enhancing the precision of weather forecasts. Despite existing uncertainties over land and ice-covered surfaces, recent advances have enhanced their use. This study examines the impact of assuming either Lambertian or specular surface reflection on the [...] Read more.

The utilisation of microwave radiances is crucial for enhancing the precision of weather forecasts. Despite existing uncertainties over land and ice-covered surfaces, recent advances have enhanced their use. This study examines the impact of assuming either Lambertian or specular surface reflection on the simulation of brightness temperatures for surface-sensitive, clear-sky AMSU-A microwave radiances across land and snow-covered areas. It represents the preliminary work before running a full assimilation and forecast impact study. Using the high-resolution HARMONIE-AROME regional modelling system, experiments were conducted to retrieve and analyse the retrieved emissivity in different conditions/seasons. The emissivity was also used as input to the radiative transfer model to simulate brightness temperatures of surface-sensitive sounding channels. The results show that the Lambertian assumption produces higher variability in dynamic surface emissivity, while the specular approach yields smaller and more consistent deviations. During winter, specular reflection shows higher first-guess departures (e.g., observations minus simulations) for surface-sensitive sounding observations, whereas in summer it performs better over land surfaces. Over snow-covered regions, the use of the Lambertian reflection to simulate the brightness temperature gives smaller mean errors for AMSU-A channels 4 (52.8 GHz) and 5 (53.59 GHz). These findings encourage further investigation into implementing a parameter that accounts for the Lambertian component of surface reflection when simulating brightness temperature in high-resolution limited-area models. Additionally, these findings provide practical guidance for configuring complex Nordic surface regional models and for future Arctic Weather Satellite microwave radiance assimilation. Full article

(This article belongs to the Special Issue Applications and New Trends in Metrology for Radar/LiDAR-Based Systems (Third Edition))

► Show Figures

Figure 1

21 pages, 23062 KB

Open AccessArticle

Does Immersive VR Alter Landscape Perception? A Comparative Evaluation of UAV-Derived VR Versus 2D Imagery in Rural Villages

by Siya Zhao, Litao Zhu, Luyi Wang, Wenzheng Jia, Hao Wang, He Wu, Bo Wang and Wen Dai

Remote Sens. 2026, 18(16), 2818; https://doi.org/10.3390/rs18162818 - 20 Aug 2026

Viewed by 343

Abstract

Traditional rural landscape evaluations have generally relied on ground-level photographs or videos. However, these approaches have limitations in spatial continuity, depth cues, and interactivity. Unmanned Aerial Vehicle (UAV) photogrammetry and immersive virtual reality (VR) were integrated into a comparative rural landscape evaluation framework [...] Read more.

Traditional rural landscape evaluations have generally relied on ground-level photographs or videos. However, these approaches have limitations in spatial continuity, depth cues, and interactivity. Unmanned Aerial Vehicle (UAV) photogrammetry and immersive virtual reality (VR) were integrated into a comparative rural landscape evaluation framework to assess landscape aesthetic quality. UAV-derived 3D village models were generated and deployed on PICO 4 headsets through Unity 3D and the Cesium plugin, providing evaluators with spatially continuous and 6DoF-enabled immersive representations of village scenes. The evaluation included ten landscape feature factors, including color harmony, vegetation richness, building layout harmony, openness of view, and sense of spatial depth. Ratings were collected from 75 valid participants across 17 villages, with village-level mean scores serving as the primary unit of inference. Paired-samples t-tests, subgroup sensitivity analysis, expert-only presentation-order sensitivity analysis, Pearson correlations, Steiger tests for dependent correlations, stepwise multiple linear regression, nested leave-one-village-out cross-validation (LOOCV), and bootstrap variable-selection stability analysis were conducted to examine differences between the 2D photo-based and VR-based conditions. The results showed that: (1) overall satisfaction was significantly higher in the VR-based condition than in the 2D photo-based condition (3.46 vs. 3.24); (2) the condition-specific regression models retained different landscape feature factors: sense of spatial depth and color harmony in the 2D photo-based model, and vegetation distribution pattern and environmental comfort in the VR-based model; and (3) the VR-based regression model had a higher condition-specific internal

R^{2}

than the 2D photo-based model (

R^{2} = 0.784

vs. 0.569). Within the present dataset, the VR-based model also showed lower SD-normalized prediction error under nested LOOCV, while bootstrap resampling showed higher selection frequencies for the predictors retained in the VR-based model. Overall, the findings demonstrate the potential of UAV-derived immersive VR for rural landscape evaluation and provide new evidence on how presentation conditions influence landscape perception and evaluation. Full article

(This article belongs to the Special Issue New Tools or Trends for Large-Scale Mapping and 3D Modelling (Second Edition))

► Show Figures

Figure 1

22 pages, 65601 KB

Open AccessArticle

Dual-Domain Illumination Prior for Low-Light Remote Sensing Image Enhancement

by Chao Wang, Zhe Pan, Liangtian He, Jun Liu, Lin Mei, Rongsheng Lin, Hongming Chen and Chuansheng Yang

Remote Sens. 2026, 18(16), 2817; https://doi.org/10.3390/rs18162817 - 20 Aug 2026

Viewed by 313

Abstract

Low-light conditions degrade remote sensing imagery by reducing contrast, distorting color, and obscuring fine terrain structures and small objects critical for Earth observation. Accurate illumination adjustment under spatially varying scene content remains challenging for existing enhancement methods, and many prior-guided approaches operate exclusively [...] Read more.

Low-light conditions degrade remote sensing imagery by reducing contrast, distorting color, and obscuring fine terrain structures and small objects critical for Earth observation. Accurate illumination adjustment under spatially varying scene content remains challenging for existing enhancement methods, and many prior-guided approaches operate exclusively in either the spatial domain or the frequency domain. In this work, we propose a Dual-Domain Illumination Prior (DDIP), a trainable dual-domain illumination-prior module that is jointly optimized with each host backbone and exploits frequency-domain and spatial-domain illumination statistics. DDIP comprises three components: a Frequency-Domain Illumination Distribution Prior (FIDP) that performs per-color-channel amplitude calibration in Fourier space to improve global brightness; a Spatial-Domain Illumination Distribution Prior (SIDP), adapted from IDP-Net, that performs multi-scale sub-region statistical correction for local illumination adjustment; and a Selective Core Feature Fusion (SCFF) module that adaptively combines the frequency-domain output, the spatial-domain output, and the original input through an attention-based gating mechanism with dual pooling. DDIP is integrated with each host backbone while leaving its main restoration blocks unchanged. In the controlled reconstruction comparisons on iSAID-dark and the evaluated general low-light benchmarks, equipping the tested backbone networks with DDIP improves PSNR and SSIM over their corresponding baselines. Complementary LPIPS and CIELAB lightness measurements characterize perceptual similarity and lightness behavior, while a fixed-detector object-detection evaluation on the tested high-resolution iSAID-dark scenes examines the effect of the enhancement pipelines under the reported synthetic low-light conditions. The ablation studies further examine the contribution of the module components within the reported experimental settings. Full article

(This article belongs to the Special Issue Hyperspectral Remote Sensing Image Analysis via Advanced Deep Learning and Computer Vision)

► Show Figures

Figure 1

27 pages, 18959 KB

Open AccessArticle

Construction and Validation of a High-Fidelity Virtual Scene for Low-Stature and High-Biodiversity Ecosystems—Simulating Multi-Modal Sensing Approaches

by Manisha Das Chaity, Ramesh Bhatta, Byron Eng and Jan van Aardt

Remote Sens. 2026, 18(16), 2816; https://doi.org/10.3390/rs18162816 - 20 Aug 2026

Viewed by 442

Abstract

The Greater Cape Floristic Region (GCFR) in South Africa is a fire-prone biodiversity hotspot where high species richness, structural complexity, and small plant sizes (0.0001–4 m²) pose substantial challenges for remote sensing-based biodiversity assessment. Spectral similarity among species and the mismatch [...] Read more.

The Greater Cape Floristic Region (GCFR) in South Africa is a fire-prone biodiversity hotspot where high species richness, structural complexity, and small plant sizes (0.0001–4 m²) pose substantial challenges for remote sensing-based biodiversity assessment. Spectral similarity among species and the mismatch between plant size and sensor pixel dimensions limit the capacity of current and forthcoming spaceborne systems to resolve individual species and accurately detect plot-level diversity changes. We therefore developed a physics-based simulation framework that couples fynbos trait measurements with radiative transfer modeling in the DIRSIG (Digital Imaging and Remote Sensing Image Generation) environment towards quantifying information loss across spectral and spatial scales and to define theoretical limits for biodiversity monitoring. We constructed a three-dimensional virtual scene of post-fire fynbos communities in Grootbos Private Nature Reserve, integrating high-resolution imagery, terrestrial laser scanning (TLS), and structure-from-motion (SfM)-derived point clouds. Field measurements of mean diameter and percent cover were used to scale vegetation models and constrain species abundance. We distributed plant instances using a blue noise sampling algorithm, guided by density maps derived from unmanned aerial system (UAS) imagery. Species-specific optical properties were parameterized using field-measured reflectance data and the PROSPECT radiative transfer model, while terrain structure was derived from SfM-based digital terrain models. The integrated scene was used to simulate multispectral (DJI Mavic 3 MSI), hyperspectral (AVIRIS-NG), and light detection and ranging (LiDAR) observations. Agreement between simulated outputs were evaluated against corresponding field-acquired datasets using spectral signatures and vegetation indices. This framework enables systematic assessment of sensor specification effects on spectral biodiversity metrics and provides a pathway for evaluating theoretical limits of species discrimination across airborne and satellite platforms. Full article

► Show Figures

Graphical abstract

Journal Menu

Journal Browser

Remote Sens., Volume 18, Issue 16 (August-2 2026) – 202 articles

Further Information

Guidelines

MDPI Initiatives

Follow MDPI