Search Results (110)

Accurate extraction of vegetation information from high-resolution remote sensing (RS) imagery is crucial for efficient urban ecological environment monitoring and land use management. However, due to the high cost of manual annotation in remote sensing imagery and the complex textural variations and spectral confusion exhibited by vegetation under different terrains and lighting conditions, precise vegetation segmentation under small-sample conditions remains a significant challenge. Using the Nanjing Zijinshan region as a case study, this research conducts a systematic comparison of eight representative models within a unified high-resolution remote sensing small-sample experimental framework to address these complexity challenges. We fine-tuned and systematically compared the recently prominent “Segment Anything Model” (SAM) series (including SAM2-Tiny, SAM2.1-Tiny, MobileSAM, and MobileSAMV2), along with classic fully supervised models (U-Net, DeepLabV3+), open-vocabulary segmentation models (SegEarth-OV), and instance segmentation models (YOLO11s-seg), helping clarify the performance boundaries and applicable conditions of different technical paradigms in vegetation segmentation. Experimental results highlight the distinctive performance characteristics of these models. Notably, fine-tuned vision foundation models (such as SAM2.1-Tiny and SAM2-Tiny) demonstrated superior segmentation performance and cross-dataset generalization capabilities, with SAM2.1-Tiny achieving the highest mean Intersection over Union (mIoU; 0.7821) on the Zijinshan dataset, a 5.5% improvement over the classic U-Net model; SAM2-Tiny also maintained the most stable generalization performance in cross-dataset testing on LoveDA, Potsdam, and Vaihingen. In contrast, zero-shot SegEarth-OV and instance segmentation model YOLO11s-seg showed relatively lower performance in current semantic segmentation tasks, revealing the application boundaries of different paradigms. Beyond these findings, to further leverage unlabeled temporal imagery and break through small-sample constraints, we propose an innovative Cross-Temporal Pseudo-Label Self-Training (CT-PLST) strategy, which successfully improved SAM2-Tiny’s mIoU from 0.7776 to 0.7888 (+1.44%), providing a low-cost efficiency enhancement solution for remote sensing segmentation under scarce annotation conditions. To promote reproducible research in remote sensing and computer vision, we publicly release the fine-tuned models, related comparative experiment code, and a high-resolution remote sensing vegetation dataset covering multi-temporal scenarios; access details are provided in the Data Availability Statement. The findings of this study, combined with the proposed CT-PLST strategy and the high-precision segmentation results achieved by vision foundation models, can strongly support tracking analysis of vegetation cover changes, urban heat island effect assessment, and exploration of ecosystem dynamic evolution. Meanwhile, these achievements also provide valuable theoretical guidance and engineering references for practitioners and researchers in finding lightweight segmentation models suitable for specific image characteristics and computational cost constraints in practical applications such as rapid disaster risk assessment or forestry resource surveys. Full article

Estimating carbon stocks in semi-urban ecosystems remains challenging due to spatial heterogeneity and the scale limitations of conventional datasets. This study aims to estimate and analyse the spatial and temporal distribution of carbon stocks at Universitas Padjadjaran using high-resolution remote sensing imagery and to support sustainable campus and green campus initiatives. Multi-temporal data from WorldView-2 (2015, 2017), WorldView-3 (2021), and Legion-03 (2025) were used to derive vegetation indices, followed by aboveground biomass (AGB) modelling through regression analysis. Carbon stock was calculated using a standard conversion factor of 0.5. The results show a consistent increase in vegetation density and carbon stock, with average values rising from 20.381 tonnes/ha in 2015 to 29.160 tonnes/ha in 2025. The use of the MSAVI produced an accurate model for predicting AGB (R² = 0.987–0.993). This study introduces a novel integration of high-resolution imagery using MSAVI to improve AGB estimation at the campus scale, providing a more detailed and reliable approach for carbon assessment in heterogeneous semi-urban environments and contributing to the implementation of sustainable, environmentally friendly campus management strategies. Full article

Tropical rainforests are characterized by high species diversity and complex canopy structure, making accurate forest stand classification important for ecosystem assessment, biodiversity monitoring, and forest carbon estimation. However, single-source remote sensing data lacks sufficient discrimination ability to address the issue of spectral similarity among classes, and conventional convolutional neural networks often struggle to extract discriminative features and integrate heterogeneous data in highly complex forests. To address these challenges, this study developed a dual-branch deep learning framework that integrates DenseNet and ConvNeXt for classification in Hainan Tropical Rainforest National Park. The framework combines sub-meter Google Earth imagery to capture spatial–textural detail with multi-temporal Sentinel-2 imagery to represent phenological variation. The results showed that multi-temporal Sentinel-2 data outperformed single-date imagery by capturing phenological patterns, and that the fusion of high-resolution spatial information and multi-temporal spectral information yielded higher accuracy than either data source alone. The dual-branch model achieved an overall accuracy of 94.47% and a Kappa coefficient of 0.94, outperforming all benchmark models. These findings indicate that branch-specific feature extraction and adaptive fusion can improve fine-scale classification in complex tropical rainforest environments. The proposed framework provides a practical approach for fine-scale forest stand mapping and may support biodiversity monitoring, ecological assessment, and sustainable forest management. Full article

Levee-like features are critical for water regulation in arid regions, but their automated extraction from remote sensing imagery remains challenging due to the scarcity of high-resolution labeled datasets. This data descriptor introduces a high-resolution remote sensing image dataset for semantic segmentation of levee-like features. The dataset covers 11 regions across Xinjiang and Gansu Province in northwestern China. It includes 459 single-phase base images with a spatial resolution of 0.50 m, as well as multi-temporal images of the same regions captured at different times. All annotations were manually drawn in polygon mode using the LabelMe tool and converted into YOLO format label files. The dataset adopts a strict strategy to prevent data leakage: first, training, validation and test sets are divided based on single-phase images, and then multi-temporal images are allocated to the corresponding data subsets according to their spatial locations. The dataset has been publicly released on the ScienceDB platform under the CC BY 4.0 license. YOLO and U-Net segmentation experiments on the test set achieved promising results, demonstrating its usability for levee-like feature segmentation. This dataset can provide fundamental data support for research on levee-like feature extraction, remote sensing change detection, and cross-region model transfer learning. Full article

Accurate crop distribution mapping using multi-temporal remote sensing has become increasingly important for agricultural monitoring and management. However, existing methods often rely on the direct stacking of multi-temporal features, which leads to feature redundancy and reduced model efficiency. To address this issue, this study proposes a winter wheat identification framework that integrates growth-stage ranking, key spectral variable selection, and cross-temporal feature fusion. Taking Shandong Province as the study area, eight typical growth stages during the 2023–2024 winter wheat growing season were analyzed using Sentinel-2 imagery. Random Forest models were first constructed for each growth stage to evaluate discriminative ability. Then, spectral variable contributions were quantified using permutation importance, and key spectral variables were selected under correlation constraints. The progressive accumulation model (PAM) was then developed according to the ranking of discriminative ability across different growth stages, while the cross-temporal fusion model (CTFM) was constructed by extracting inter-stage mean values (mean) and inter-stage differences (diff) of key variables. The results show that the feature space was reduced from 64 dimensions (8 stages × 8 variables) to 16 key variables, substantially improving feature representation efficiency. Among the eight growth stages, the jointing, overwintering, heading, and grain-filling stages exhibited relatively strong discriminative ability. In the cross-temporal experiments, CTFM M6, which integrates information from the top six growth stages ranked by discriminative ability, achieved an overall accuracy (OA) of 0.9658 and a user’s accuracy (UA) of 0.9609 using only 10 fused features, providing the best balance between identification accuracy and feature dimensionality. Based on this model, a 10 m resolution winter wheat distribution map of Shandong Province was generated, and the estimated planting area showed high consistency with statistical yearbook data. These results demonstrate that the proposed strategy can effectively reduce feature dimensionality while maintaining high identification accuracy, providing an efficient and scalable approach for regional-scale remote sensing mapping of winter wheat. Full article

Accurate yield estimation is vital for precision wheat management and breeding. Traditional methods based on single growth stages or single-source data cannot capture cumulative growth effects, limiting prediction accuracy. UAV remote sensing provides high-resolution, multi-source, and multi-temporal data, enabling improved non-destructive yield estimation. In this study, UAV-based multispectral and RGB imagery were collected at six key growth stages, and vegetation indices, texture, and color features were extracted to develop yield prediction models using RF, XGBoost, and KNN under single- and multi-temporal scenarios. The results showed that red-edge-based vegetation indices were highly sensitive to wheat yield and outperformed texture- and color-based features. Multi-feature fusion further improved prediction accuracy at key growth stages, particularly during booting and flowering (R² = 0.53–0.67). Compared with single-temporal models, multi-temporal data fusion significantly enhanced yield estimation accuracy, achieving a maximum R² of 0.72 by integrating data from the late-jointing, booting and flowering stages. Among the algorithms, XGBoost and KNN exhibited superior accuracy and stability across most growth stages. Overall, these results demonstrate that integrating UAV-based multi-source and multi-temporal remote sensing data effectively improves the accuracy and robustness of wheat yield estimation, providing valuable technical support for precision agriculture and phenotyping-assisted breeding. Full article

The detection and monitoring of asbestos–cement roofing remain a critical public health and environmental challenge, especially in urban and suburban areas where asbestos-containing materials are still widespread due to their extensive use in the 20th century. Although hyperspectral and high-resolution multispectral remote sensing have proven effective for mapping asbestos–cement roofs, many existing approaches rely on proprietary software, limiting transparency, reproducibility, and large-scale adoption. This study presents a fully reproducible, cost-free Python-based workflow for the detection and temporal monitoring of asbestos–cement roofing using high-resolution multispectral WorldView-3 imagery. The workflow integrates atmospheric correction (using the Py6S radiative transfer model), spatial preprocessing, supervised pixel-based classification, postprocessing, and building-level aggregation within an open framework. A Maximum Likelihood Classifier is applied to VNIR and SWIR data using empirically defined roof typologies to enhance class separability. Pixel-level results are aggregated to the building scale through adaptive thresholding enabling the translation of spectral classifications into meaningful building-level information. Tested over the city of Mantua (Italy), the approach achieved reliable classification performance and enabled multi-temporal comparison to identify changes potentially due to roof remediation. Evaluation metrics (precision, recall, and F1-score) highlight the importance of carefully choosing the building-level threshold. By relying exclusively on open-source tools, the workflow enhances transparency, reproducibility, and scalability for long-term monitoring. Full article

Urban agriculture plays a critical yet increasingly complex role in sustainable urban development, especially in high-density regions undergoing rapid transformation. Accurate mapping of its spatial distribution and functional composition remains a methodological challenge due to its fragmented landscape, small plot sizes, and multifunctional nature. This study addresses this gap by developing and applying a novel hierarchical classification framework that integrates agricultural land cover types with key socio-economic functions to map urban agriculture in the Pearl River Delta (PRD), China. This framework is structured around agricultural land categories (i.e., cropland, garden, forest, grass, and water body) and further delineated by two primary production functions, planting and breeding, with a third functional dimension, leisure activities, proposed as a conceptual extension for future research. Using unmanned aerial vehicle (UAV) imagery and high-resolution satellite data, we constructed a spatial sample database for urban agriculture. The random forest algorithm was applied to classify urban agriculture with Gaofen-2 imagery, generating detailed spatial distribution maps across the study area, with consistently reliable overall accuracy (79.07–81.82%), though this may be slightly optimistic due to potential spatial autocorrelation between training and testing samples. While the framework performed exceptionally well for spectrally and spatially distinct classes such as water bodies and perennial plantations, challenges remained in discriminating among annual field crops due to spectral similarity. These findings underscore the potential of integrating multi-temporal remote sensing data to capture phenological variations for improved classification. This study provides a replicable, functionally informed mapping approach that not only advances the methodological toolkit for urban agriculture characterization but also offers a valuable evidence base for land use planning, agricultural policy, and sustainable urban development in rapidly urbanizing regions. Full article

Timely and accurate mapping of tree species is essential for forest resource inventory, biodiversity conservation, and sustainable ecosystem management, particularly in dryland environments where structural heterogeneity, spectral similarity, and data scarcity complicate classification. This study develops a machine learning-based framework implemented in Google Earth Engine to map dominant tree species in the Elnour Natural Forest Reserve (ENFR), Blue Nile, Sudan, using multi-temporal and multi-sensor remote sensing data. Multi-temporal Landsat 5 TM, Landsat 8 OLI, and Sentinel-2 MSI imagery were integrated with vegetation index (NDVI), topographic variables derived from a digital elevation model (DEM), and field observations. The performance of Random Forest (RF), Support Vector Machine (SVM), Classification and Regression Trees (CART), and an unweighted ensemble approach was evaluated across four reference years (2008, 2013, 2018, and 2021). Results show that RF and SVM consistently achieved high classification performance, with overall accuracy (OA) ranging from 85.0% to 92.0% and Kappa coefficients (κ) from 0.81 to 0.89, while maintaining stable and ecologically realistic species-area estimates. CART showed greater sensitivity to class imbalance and overestimated minor species (OA = 72.0–80.0%, κ = 0.65–0.74), whereas the ensemble approach amplified misclassification of rare classes (OA = 78.0–84.0%, κ = 0.70–0.78). The integration of Sentinel-2 data improved species discrimination due to enhanced spatial and spectral resolution, particularly in the red-edge region; however, algorithm selection remained the dominant factor controlling performance. Feature importance analysis identified near-infrared (NIR), shortwave infrared (SWIR), and NDVI variables as the most influential predictors. Multi-temporal analysis revealed declining class separability, reflected by decreasing MCC values, and a shift in species composition, including a decline in Acacia seyal (Delile) and an increase in Sterculia setigera Delile. These patterns indicate increasing ecological complexity driven primarily by anthropogenic pressures, with climatic variability acting as an additional stressor. Full article

Accurate exploitation of spatial structures and spectral characteristics is essential for fine-grained crop classification using remote sensing imagery. Although multi-source remote sensing data provide complementary information, most existing methods implicitly assume homogeneous data sources with consistent spatial resolution. In practice, high spatial resolution and rich spectral information are usually provided by different sensors, making cross-source spatial–spectral fusion a non-trivial challenge. To address this issue, we propose SSF-TransUnet, a dual-branch spatial–spectral joint modeling framework for fine crop classification. The proposed network explicitly decouples spatial structure extraction and spectral discriminability learning by jointly utilizing high spatial resolution imagery and multi-spectral observations acquired from different satellite sensors within a unified architecture. To support model training and evaluation, we construct SSCR-Agri, a spatial–spectral complementary resolution agricultural dataset integrating meter-level GF-2 imagery and multi-spectral Sentinel-2 data from five representative agricultural regions in northern China, covering five crop categories including corn, rice, wheat, potato, and others. Extensive experiments demonstrate that SSF-TransUnet consistently outperforms representative CNN-based and hybrid CNN–Transformer models. The proposed method achieves an overall accuracy (OA) of 81.84% and a mean Intersection over Union (mIoU) of 0.6954 in fine-grained crop classification, effectively distinguishing crops. These results highlight the effectiveness of spatial–spectral joint modeling for high-resolution crop mapping and demonstrate its potential for precision agriculture and large-scale agricultural monitoring applications, and shows a promising mechanism when combined with multi-temporal observations. Full article

Forest fires occur frequently in China; however, the complex terrain and incomplete road networks severely constrain ground rescue efficiency. Accurate forest road information is essential for the optimization of emergency response and rescue force deployment. Existing road extraction algorithms are primarily designed for urban environments and exhibit limited efficacy in forest scenarios due to dense canopy, complex background interference and specific forest road features. To address this gap, this study proposes a forest road extraction method based on an enhanced DeepLabv3+ model using multi-temporal, high-resolution satellite imagery. Specifically, a Multi-Scale Channel Attention (MCSA) mechanism is embedded in skip connections to suppress background interference, while strip pooling is integrated into the Atrous Spatial Pyramid Pooling (ASPP) module to better capture slender road features. A composite Focal-Dice loss function is also constructed to mitigate sample imbalance. Finally, by applying the model in multi-temporal remote sensing images, a fusion strategy is introduced to integrate multi-seasonal road masks to enhance overall accuracy and topological integrity. Experimental results show that the proposed method achieves a precision of 54.1%, an F1-Score of 59.3%, and an IoU of 41.8%, effectively enhancing road continuity and providing robust technical support for fire-rescue decision-making. Full article

Global climate change is increasing extreme rainfall events, and severe floods are becoming more frequent. Flood storage and detention basins (FSDBs) are an important part of the flood control system in China. They play a key role in regional flood emergency response and regulation. Therefore, accurate simulation of flood evolution after the activation of FSDBs is urgently needed. This study proposes a high-accuracy flood evolution simulation method that combines terrain clustering and physical propagation constraints. We first build a 2 m resolution digital elevation model (DEM) using GF-7 stereo imagery and laser altimetry data. We then introduce an improved superpixel segmentation algorithm (TSLIC). This method reduces the number of computational units while preserving key micro-topographic features. It groups high-resolution grids into terrain units with similar elevation characteristics and continuous spatial structure. Based on these terrain units, we develop a flood evolution model called RS-CFPM. The model combines flow velocity estimated from the Manning equation with flood propagation speed derived from radar remote sensing. It uses a water balance framework and includes a propagation time delay constraint. This design helps overcome the limitation of traditional static inundation methods that ignore flood travel time. We apply the proposed method to simulate the flood inundation process during the “23·7” extreme basin-scale flood event in the Haihe River Basin. Comparison with multi-temporal radar observations shows that the errors of simulated water level and inundation extent in the Dongdian FSDB are both within 10%. The computational efficiency is also improved by more than 60% compared with traditional methods. This study provides a new approach for rapid and accurate simulation of flood inundation processes in FSDBs under emergency conditions. The method can support flood emergency operation and decision-making. Full article

High-spatial-resolution remote sensing imagery provides a data foundation for fine-grained land use classification. However, due to long revisit cycles and susceptibility to cloud cover, large-area imagery often suffers from temporal inconsistency, which severely limits the classification accuracy of traditional unified models. To address this issue, this study proposes a geographic entity-oriented, spatiotemporally coupled land use classification method for high-resolution remote sensing imagery, with agricultural land (including paddy fields, dry farmland and gardens) as an example for validation. In this method, the study area is first divided into multiple sub-regions based on image acquisition time, ensuring temporal consistency within each sub-region. A dedicated deep texture feature extraction model is then constructed for each sub-region. This model is adapted from the advanced CAPTN texture recognition network: its classification head is removed, and a multi-scale feature fusion module is introduced, transforming it into an encoder focused on extracting spatial texture feature maps. Additionally, a self-supervised loss function combining masked feature reconstruction and cross-view consistency is designed to improve the quality of the learned texture features. During the prediction stage, the corresponding feature extractor is invoked based on the temporal phase of the imagery to generate a full-region texture feature map. This feature map is then cropped using land parcel vectors, and statistical feature vectors describing the texture attributes of each parcel are formed by calculating the mean and standard deviation of the features within each parcel. Finally, a Random Forest classifier is employed to determine the land parcel categories. This study uses the Jiangjin District of Chongqing City as the experimental area. The results show that, compared to training a unified deep learning model directly on full-region multi-temporal imagery or using traditional texture features, the proposed spatiotemporally coupled classification framework achieves significant improvements in overall accuracy and Kappa coefficient, reaching 92.3% and 0.89, respectively. Full article

Deep learning-based methods for land cover semantic change detection utilizing high-resolution, multi-temporal remote sensing imagery have emerged as a research hotspot. However, traditional CNN methods often struggle to preserve long-range spatial context information and face challenges in detecting land cover types with complex semantic change patterns in natural scenes. To address these issues, this study proposes a novel network architecture that integrates a Siamese network with differential structures and a Transformer. First, we introduce residual learning modules to improve the extraction of differential features and strengthen the representation of local features. Second, we integrate the Cross-Shaped Window (CSWin) Transformer into a differential-based Siamese network to enhance global feature extraction. To promote model training and evaluation, we propose a rural land cover change detection dataset—a high-precision dataset comprising 6 main rural land cover types. Ablation and comparative experiments were conducted on the publicly available SECOND datasets and the self-built RLCD dataset. Ablation studies on the RLCD dataset demonstrate that DSTNet achieves significant improvements over the baseline, with increases of 1.77%, 1.95%, 2.57%, and 0.92% in mIoU, Sek, Fscd, and OA. Comparative experiments on the SECOND datasets reveal that the mIoU, Sek, Fsd, and OA scores of DSTNet surpassed the second-best accuracy by 1.04%, 2.15%, 2.28%, and 0.72%. Full article

The Damietta–Port Said coast, Nile Delta, has experienced extreme morphological change over the past four decades due to sediment reduction due to Aswan High Dam and continued anthropogenic pressures. Using multi-temporal Landsat (1985–2025) and high-resolution RapidEye and PlanetScope imagery with 50 m-spaced transects, the study documents major shoreline shifts: the Damietta sand spit retreated by >1 km at its proximal apex while its distal tip advanced by ≈3.1 km southeastward under persistent longshore drift. Sectoral analyses reveal typical structure-induced patterns of updrift accretion (+180 to +210 m) and downdrift erosion (−50 to −330 m). To improve predictive capability beyond linear DSAS extrapolation, Nonlinear Autoregressive Exogenous (NARX) and Bidirectional Long Short-Term Memory (BiLSTM) neural networks were applied to forecast the 2050 shoreline. BiLSTM demonstrated superior stability, capturing nonlinear sediment transport patterns where NARX produced unstable over-predictions. Furthermore, coupled wave–flow modeling validates a sustainable management strategy employing successive short groins (45–50 m length, 150 m spacing). Simulations indicate that this configuration reduces longshore current velocities by 40–60% and suppresses rip-current eddies, offering a sediment-compatible alternative to conventional breakwaters and seawalls. This integrated remote sensing, hydrodynamic, and AI-based framework provides a robust scientific basis for adaptive, sediment-compatible shoreline management, supporting the long-term resilience of one of Egypt’s most vulnerable deltaic coasts under accelerating climatic and anthropogenic pressures. Full article