IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing最新文献

{"title":"Road Crack Detection and Orientation Estimation Using Airborne Synthetic Aperture Radar","authors":"Arun Babu;Stefan V. Baumgartner;Gerhard Krieger","doi":"10.1109/JSTARS.2025.3613166","DOIUrl":"https://doi.org/10.1109/JSTARS.2025.3613166","url":null,"abstract":"Cracks on road surfaces are a significant safety hazard that can progress into larger potholes, posing risks to vehicles and passengers. Synthetic aperture radar (SAR) data acquired by high-resolution airborne SAR systems are sensitive to changes on the road surface and can be utilized for periodic road condition monitoring. This study proposes a novel method that combines an adaptive thresholding algorithm with the Radon transform for detecting road cracks and estimating both their severity and orientation. In this approach, the adaptive thresholding algorithm detects the cracks, while the Radon transform qualitatively quantifies their severity using the maximum Radon magnitude from the sinogram and estimates their orientation as bearing angles relative to true north. While the proposed method is applicable to various airborne SAR platforms, it is demonstrated in this study with X-band airborne SAR data acquired by DLR’s F-SAR system with a spatial resolution of 25 cm. The detected cracks and orientations were validated against Google Earth images, showing close agreement with the locations and orientations of the actual cracks. This research underscores the potential of airborne SAR data in supporting predictive road maintenance efforts through early identification of surface defects.","PeriodicalId":13116,"journal":{"name":"IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing","volume":"18 ","pages":"24883-24895"},"PeriodicalIF":5.3,"publicationDate":"2025-09-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11175491","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145210167","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"3D Change Detection of Urban Vegetation Using Integrated TLS and UAV Photogrammetry Point Clouds","authors":"Osama Bin Shafaat;Heikki Kauhanen;Arttu Julin;Matti T. Vaaja","doi":"10.1109/JSTARS.2025.3612739","DOIUrl":"https://doi.org/10.1109/JSTARS.2025.3612739","url":null,"abstract":"Urbanization has brought about notable transformations in urban green areas within cities, affecting both environmental quality and the well-being of inhabitants. As a result, it is essential to monitor variations in urban vegetation through remote sensing methods. This research aims to overcome the shortcomings of conventional remote sensing approaches by integrating terrestrial laser scanning (TLS) with UAV-based photogrammetry for effective vegetation monitoring using change detection methods. For instance, the traditional remote sensing limitations include cloud coverage in remote sensing images, illumination issues, vertical shadows, and sensor-specific issues such as geometric and radiometric distortions that restrict the spatiotemporal availability of the ground surface information and limit the change detection analysis. This research focuses on detecting changes in the Malminkartano area of Helsinki during the leaf-off and leaf-on periods of 2022. 2D point cloud data were analyzed using the Multiscale Model-to-Model Cloud Comparison algorithm. The findings demonstrate the method’s capability to identify growth in urban vegetation up to 2.8 m. Additionally, accuracy evaluations indicated that the 95% confidence interval corresponded to a difference of approximately 4 cm for both TLS and UAV photogrammetric datasets. The study highlights processing-related uncertainties, including point density, alignment, vertical accuracy, and scale variation. Addressing these sources of error in future studies is essential for reliable estimation of tree attributes.","PeriodicalId":13116,"journal":{"name":"IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing","volume":"18 ","pages":"24976-24989"},"PeriodicalIF":5.3,"publicationDate":"2025-09-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11174137","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145255863","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A Remote Sensing Detection Method for Ceracris Kiangsu Tsai in Moso Bamboo Forests Integrating SMOTE Algorithm and Multi-Model Ensemble","authors":"Na Qin;Zhanghua Xu;Wensen Yu;Shancai Xu;Rengui Cheng;Yongwei Cao;Xier Yu;Weili Liu;Xiaoyu Guo;Fengying Guan","doi":"10.1109/JSTARS.2025.3612436","DOIUrl":"https://doi.org/10.1109/JSTARS.2025.3612436","url":null,"abstract":"<italic>Ceracris kiangsu</i> Tsai (<italic>C.kiangsu</i>) is one of the main leaf-feeding pests in Moso bamboo forests. An in-depth exploration of its response mechanism is crucial for achieving large-scale, precise detection and maintaining the healthy development of Moso bamboo forests. However, existing research on <italic>C.kiangsu</i> pest detection is relatively limited, with traditional forest pest models often constrained by unbalanced samples for generalization. This study integrates field survey data with Sentinel-2 MSI data to explore the remote sensing response mechanism of the pest; proposes a collaborative detection method for <italic>C.kiangsu</i> infestations in Moso bamboo forests that combines the SMOTE algorithm with a multi-model ensemble, optimizes model parameters via the Bayesian algorithm, and simultaneously uses SHAP values to deeply analyze model interpretability and excavate pest detection indicators. The results showed that: 1) Leaf LCC, LWC, and LDMC exhibit excellent responsiveness to <italic>C.kiangsu</i> pest infestations (<italic>p</i> < 0.01), with vegetation indices outperforming moisture indices and texture features. 2) The optimal Stacking ensemble learning hybrid model achieves OA of 82.22% and Kappa of 0.7625, improving by 3.5% and 0.0468 over the best single LightGBM model. 3) SHAP analysis reveals that the vegetation index RVI is the core indicator for pest detection, ranking among the top three in feature importance in each base model. The contribution to the meta-model is LightGBM > ET > SVM > XGBoost. This study successfully breaks through the accuracy bottleneck of traditional single models, providing a solid scientific basis for Moso bamboo pest management and practical significance for its sustainable development.","PeriodicalId":13116,"journal":{"name":"IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing","volume":"18 ","pages":"25005-25023"},"PeriodicalIF":5.3,"publicationDate":"2025-09-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11174951","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145255913","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Cascaded Fine Detection of Small Object in Large-Scale SAR Images Leveraging AIS Data","authors":"Borui Li;Baoxiang Huang;He Gao;Ge Chen","doi":"10.1109/JSTARS.2025.3612035","DOIUrl":"https://doi.org/10.1109/JSTARS.2025.3612035","url":null,"abstract":"Detecting small objects in large-scale synthetic aperture radar (SAR) images presents significant challenges due to their subtle features and the interference from complex backgrounds. This is particularly problematic when identifying maritime vessels, as SAR imagery often includes extensive geographic information, with some land features resembling vessels. These similarities increase the likelihood of false detections, complicating the accurate identification and tracking of vessels. To address these challenges, this article proposes a cascade fine detection methodology for small objects in large-scale SAR images, leveraging automatic identification system (AIS) data for enhanced accuracy. Specifically, to address the issue of small object detection in large-scale scenes, we design a remote sensing detection network (RSDNet), which optimizes feature extraction and multiscale information fusion to capture more refined object details. To minimize false detections, AIS data is embedded into the detection framework as prior knowledge through cross-modal data fusion. Thus, AIS data is integrated with SAR images in both spatial and temporal dimensions. Meanwhile, we validate RSDNet’s accuracy using three publicly accessible datasets as well as a self-built dataset from a remote sensing object detection vessel (RSOD vessel). Furthermore, to assess the efficacy of our approach, we create a SAR–AIS matching dataset in the absence of a benchmark dataset for AIS and SAR matching. Finally, the extensive experimental results indicate that the proposed methodology can provide promising detection of small objects in large-scale SAR images.","PeriodicalId":13116,"journal":{"name":"IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing","volume":"18 ","pages":"25124-25138"},"PeriodicalIF":5.3,"publicationDate":"2025-09-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11173937","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145255847","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A Chlorophyll-a Concentration Inversion Model Based on Adaptive Fine-Tuning Transfer Learning Method and Self-Attention Mechanism","authors":"Jianyong Cui;Shuhang Hou;Jie Guo;Mingming Xu;Hui Sheng;Shanwei Liu;Muhammad Yasir;Ying Zhang","doi":"10.1109/JSTARS.2025.3611596","DOIUrl":"https://doi.org/10.1109/JSTARS.2025.3611596","url":null,"abstract":"Chlorophyll-a (Chl-a) is essential for assessing water quality and aquatic ecosystems. Traditional in situ monitoring suffers from sparse stations and limited samples, and machine learning often overfits under such conditions. Transfer learning offers a solution by leveraging pretrained knowledge, but domain discrepancies remain a major challenge. To improve cross-domain Chl-a inversion under few-shot settings, this study proposes GSA-ICPO-Former, which integrates a Transformer architecture, an improved crested porcupine optimizer (ICPO), and a gradient sensitivity-based adaptive (GSA) fine-tuning method. The Transformer captures global spectral dependencies through self-attention; ICPO enhances hyperparameter optimization and prevents overfitting; GSA adjusts model parameters to adapt source-domain knowledge to target-domain features. After ICPO optimization, the model’s <inline-formula><tex-math>$R^{2}$</tex-math></inline-formula> in the source domain improved from 0.73 to 0.84, and RMSE decreased from 2.39 to 1.86 <inline-formula><tex-math>$mu$</tex-math></inline-formula>g/L, representing improvements of 15.07% and 22.18%, respectively. In the target domain, compared with ICPO-Former, GSA-ICPO-Former raised <inline-formula><tex-math>$R^{2}$</tex-math></inline-formula> from 0.54 to 0.79 and reduced RMSE by 32.97%, demonstrating enhanced cross-domain adaptability. Ablation studies confirmed the effectiveness of each optimization component. The model also showed robust performance across different coastal environments, including both nearshore (Tangdao Bay) and open-sea (Rongcheng) regions. It accurately tracked Chl-a distribution with limited samples and showed promise in dynamic algal bloom monitoring. Overall, the proposed model effectively bridges domain gaps and maintains high prediction accuracy in low-sample, cross-regional scenarios, providing a practical tool for satellite-based coastal water quality monitoring.","PeriodicalId":13116,"journal":{"name":"IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing","volume":"18 ","pages":"25357-25373"},"PeriodicalIF":5.3,"publicationDate":"2025-09-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11172705","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145255848","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Improving the Spatial Continuity of GEDI Aboveground Biomass Density Products Using Multisource Remote Sensing Data With Consideration of Spatial Correlation and Heterogeneity","authors":"Wankun Min;Wenli Huang;Yumin Chen;Rui Xu;Lanhua Bao","doi":"10.1109/JSTARS.2025.3611427","DOIUrl":"https://doi.org/10.1109/JSTARS.2025.3611427","url":null,"abstract":"Accurate, long-term estimation of forest aboveground biomass density (AGBD) is essential for monitoring terrestrial ecosystem dynamics and quantitatively assessing the capacity of forests in the global carbon cycle and their contribution to mitigating climate change. Meanwhile, plot-level AGBD measurements are highly accurate but lack spatial and temporal continuity. Global ecosystem dynamics investigation (GEDI) provides spatially discontinued estimates of global-level AGBD. To overcome these challenges, we proposed a model incorporating spatial correlation and heterogeneity for AGBD estimation, which combines GEDI gridded data with multisource remote sensing data. To further mitigate the spatial discontinuities in GEDI-derived AGBD distribution, a LightGBM model (LGB_EV_SIT) incorporating spatial eigenvectors (EV) and spatial interaction terms (SIT) was proposed. Using GEDI-derived AGBD as a reference, key predicted indicators were identified and ranked based on multisource remote sensing variables. A spatial weight matrix was constructed to reflect the spatial distribution of the GEDI AGBD grids. Spatial EVs and SITs were extracted based on the spatial weight matrix. Compared with the LightGBM model (LGB) without considering the correlation and interaction (<italic>r</i> = 0.61–0.82, MRE = 38.6% –54.3%, RMSE = 45.91–54.60 Mg/ha), LGB_EV_SIT exhibits better performance (<italic>r</i> = 0.85–0.87, MRE = 31.5% –39.1%, RMSE = 36.47–46.17 Mg/ha). Furthermore, the framework was applied to map forest AGBD across the western USA from 2019 to 2022, enabling the detection of forest changes over time. The proposed model provides new possibilities to accurately generate wall-to-wall maps of forest biomass and carbon stocks over a long time series.","PeriodicalId":13116,"journal":{"name":"IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing","volume":"18 ","pages":"24783-24800"},"PeriodicalIF":5.3,"publicationDate":"2025-09-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11172296","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145210023","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Daily 9-km Soil Moisture Retrieval With Reconstructed GNSS-R Observations","authors":"Wentao Yang;Fei Guo;Xiaohong Zhang;Zhiyu Zhang;Yifan Zhu;Zheng Li;Dengkui Mei","doi":"10.1109/JSTARS.2025.3611668","DOIUrl":"https://doi.org/10.1109/JSTARS.2025.3611668","url":null,"abstract":"Spaceborne global navigation satellite system-reflectometry (GNSS-R) observations are effective for extensive monitoring of soil moisture (SM). Reliance on gridded data from other remote-sensing systems is the general way of GNSS-R SM retrieval. Increasing the grid size improves temporal coverage but decreases spatial resolution, which may mask changes in SM within the grid. Conversely, finer grids result in higher spatial resolution but less temporal resolution, leading to discontinuities in SM temporal dynamics. Therefore, GNSS-R observation gridding cannot achieve high spatiotemporal resolutions simultaneously. To leverage the potential of GNSS-R technology in SM retrieval with a high spatiotemporal resolution, this study introduced a reconstruction method for spaceborne GNSS-R observations to generate gridded observations with both high temporal and spatial resolution. This method was used to build a pixel-by-pixel multivariate temporal fitting model that incorporates the characteristics of GNSS-R observations and the associated driving factors. Specifically, this study utilizes cyclone GNSS (CYGNSS) sliding average observations to supplement daily trend observations and integrates ancillary data from other satellite missions as adjustments to daily variability, which generates spatiotemporally seamless CYGNSS observations. The reconstructed CYGNSS observations were then utilized to provide daily quasi-global SM retrievals on a 9-km grid. The root-mean-square error (RMSE) of the SM retrievals was 0.052 <inline-formula><tex-math>$mathrm{c}{{mathrm{m}}^{3}}{rm{/c}}{{mathrm{m}}^{3}}$</tex-math></inline-formula>, which are comparable with the accuracy of the SM retrievals conducted prior to reconstruction. Similarly, independent evaluations of local in situ sites demonstrated that the RMSE and correlation (<italic>R</i>) of the SM retrieval were 0.051 <inline-formula><tex-math>$mathrm{c}{{mathrm{m}}^{3}}{rm{/c}}{{mathrm{m}}^{3}}$</tex-math></inline-formula>and 0.77, respectively. The SM from the reconstructed CYGNSS observations also captured local wet and dry dynamics, which are consistent with the performance of the SM retrieved from the original CYGNSS observations. Notably, the temporal resolution was improved by 258% over the original CYGNSS observations at 9 km. Therefore, we argue that the method used in this study addresses the problem of mutual constraints between spatial and temporal resolutions in quasi-global CYGNSS SM retrieval without the loss of SM retrieval accuracy. Furthermore, the observation reconstruction method developed in this study may be a promising reference for monitoring other geophysical parameters with high spatiotemporal resolution in the GNSS-R domain.","PeriodicalId":13116,"journal":{"name":"IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing","volume":"18 ","pages":"25039-25053"},"PeriodicalIF":5.3,"publicationDate":"2025-09-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11172358","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145255836","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Shadow-Aware Moving Target Detection in ViSAR: A Multiscale CNN–Transformer Hybrid Detection Framework","authors":"Shangqu Yan;Yaowen Fu;Ruofeng Yu;Chenyang Luo;Wenpeng Zhang;Wei Yang","doi":"10.1109/JSTARS.2025.3611080","DOIUrl":"https://doi.org/10.1109/JSTARS.2025.3611080","url":null,"abstract":"Based on the all-weather imaging capability of synthetic aperture radar (SAR), video synthetic aperture radar (ViSAR) enables dynamic ground target monitoring through high-frame-rate continuous observation. Since moving targets’ shadows remain stable and unaffected by defocusing in ViSAR images, this property is widely leveraged for ViSAR moving target detection. However, ViSAR moving target detection faces challenges including low contrast in moving targets’ shadows, complex background clutter interference, and the difficulty of simultaneously detecting small and medium-sized shadow targets. Existing deep learning-based detection methods for ViSAR struggle to address these challenges effectively. To overcome these limitations, we propose a multiscale convolutional neural network (CNN)–Transformer hybrid detection framework, which consists of three key components. First, in the preprocessing stage of ViSAR images, an improved low-rank representation algorithm is used to suppress background clutter and stationary targets’ shadows, and enhance the contrast of moving targets’ shadows. Second, the CNN’s feature representation capability for small and medium-sized shadow targets is enhanced by improving the backbone and designing a novel feature pyramid network structure. This addresses the challenge of detecting small and medium-sized shadow targets simultaneously. Finally, within the Transformer architecture, a novel proposal generation module and a contrastive denoising training strategy are integrated into the Deformable Transformer to mitigate ambiguous semantic encoding and accelerate convergence in DETR-like detectors. Experimental results on the real-world ViSAR dataset demonstrate that the proposed framework achieves state-of-the-art performance.","PeriodicalId":13116,"journal":{"name":"IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing","volume":"18 ","pages":"24801-24815"},"PeriodicalIF":5.3,"publicationDate":"2025-09-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11168256","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145210068","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"FADet: A Frequency-Aware Detection Framework for Infrared Small Target Detection","authors":"Dan Feng;Jian Xu;Ke Li;Zhi Ma;Wen Li;Di Wang","doi":"10.1109/JSTARS.2025.3611492","DOIUrl":"https://doi.org/10.1109/JSTARS.2025.3611492","url":null,"abstract":"Infrared small target detection (IRSTD) remains a critical yet highly challenging task in the field of object detection. Due to the tiny target size and the absence of rich texture information, general-purpose detectors often suffer substantial performance degradation when applied to this task. This performance degradation is mainly due to their limited ability to extract discriminative features, resulting in frequent missed detections and false alarms that compromise the reliability of detection systems. To address these challenges, we propose <bold>FADet</b>, a novel <bold>F</b>requency-<bold>A</b>ware <bold>Det</b>ection framework specifically designed to capture the unique representational characteristics of small targets. Specifically, we introduce a Frequency-Guided Visual Encoder that leverages the Haar Wavelet Transform to explicitly decompose spatial features into high- and low-frequency components. An attention mask is then derived from the high-frequency components to selectively preserve informative fine-grained details. This process effectively alleviates the over-smoothing effect typically induced by convolutional operations, thereby significantly enhancing the saliency of small targets in the detection framework. Furthermore, we propose a Multiscale Feature Gather-Distribute module that aggregates multiscale semantic cues and redistributes them across different feature hierarchies, thereby enabling more effective feature interaction and fusion. Extensive experiments on three public benchmark datasets (e.g., NUAA-SIRST, NUDT-SIRST, and IRSTD-1 K) demonstrate that FADet achieves superior performance, setting new state-of-the-art results in infrared small target detection.","PeriodicalId":13116,"journal":{"name":"IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing","volume":"18 ","pages":"24963-24975"},"PeriodicalIF":5.3,"publicationDate":"2025-09-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11169394","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145255876","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Divide-and-Conquer: Frequency-Domain Decoupling Strategy for Low-Light Remote Sensing Image Enhancement","authors":"Bosong Zhuang;Zishu Yao;Guodong Fan;Jinjiang Li","doi":"10.1109/JSTARS.2025.3611820","DOIUrl":"https://doi.org/10.1109/JSTARS.2025.3611820","url":null,"abstract":"Low-light remote sensing (RS) images typically cover vast areas. They contain objects of various scales and have localized light sources. This makes it challenging to enhance brightness while preserving fine image structures. Existing approaches are primarily designed in the spatial domain. However, due to the tight coupling between illumination degradation and structural information, these methods often struggle to achieve effective enhancement. In this article, we propose a divide-and-conquer frequency domain decoupling enhancement strategy. Specifically, by exploring the decoupling properties of the frequency domain, we design a <italic>light contrastive regularization</i> that constrains the model to focus solely on brightness distribution in the contrastive space while reducing interference from redundant information. In addition, we introduce a novel <italic>phase mamba enhancement network</i>, which leverages the unique continuity of the frequency domain. By employing a continuous scanning mechanism, our model effectively captures long-range dependencies in low-light RS images, enabling finer grained structural restoration. Extensive experiments demonstrate that our method surpasses state-of-the-art approaches both qualitatively and quantitatively.","PeriodicalId":13116,"journal":{"name":"IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing","volume":"18 ","pages":"24936-24946"},"PeriodicalIF":5.3,"publicationDate":"2025-09-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11169498","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145255844","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}