ISPRS Journal of Photogrammetry and Remote Sensing最新文献

{"title":"Gated multi-source fusion with geometric sequence modeling for novel urban structure discovery","authors":"Jing Du, John Zelek, Dedong Zhang, Jonathan Li","doi":"10.1016/j.isprsjprs.2025.09.017","DOIUrl":"https://doi.org/10.1016/j.isprsjprs.2025.09.017","url":null,"abstract":"Novel Class Discovery (NCD) in 3D point cloud semantic segmentation presents critical challenges for urban management systems, where models must segment previously unseen object classes in rapidly evolving urban environments. Traditional 3D semantic segmentation models struggle to adapt to heterogeneous spatial characteristics and complex geometric structures of urban point clouds, limiting their ability to handle novel objects without extensive retraining. This paper introduces Adaptive Geometric Discovery Network (AGDNet), a comprehensive framework enhancing NCD through three key innovations: Adaptive Geometric Sequence Modeling module (AGSM), Dynamic Gaussian Embedding module (DGE), and Gated Multi-Source Feature Fusion module (GMSFF). AGSM addresses heterogeneous spatial characteristics through density-aware adaptive sampling, dynamic grouping, and multi-aspect geometric feature encoding. DGE represents point clouds as learnable 3D Gaussians parameterized by position, scale, orientation, and features, providing continuous probabilistic representations capturing both local geometric details and global spatial contexts. GMSFF integrates features from AGSM, DGE, and MinkowskiNet through context-aware gating mechanisms. The framework introduces three specialized knowledge transfer objectives for NCD: Prototype Relation Loss establishes semantic connections between known and novel class prototypes; Contrastive Alignment Loss creates instance-level semantic bridges; and Semantic Transfer Loss enables distribution-based knowledge propagation. These objectives bridge the semantic gap between known and novel categories while mitigating class imbalance challenges. Comprehensive evaluation on Toronto-3D, SemanticSTF, and SemanticPOSS datasets demonstrates significant improvements over state-of-the-art methods NOPS and CHNCD. For novel class discovery, the framework achieves average improvements of 6.47%/3.48%, 4.61%/3.12%, and 6.64%/4.24% in novel class mean Intersection over Union (mIoU) over NOPS/CHNCD respectively. For overall performance, improvements reach 6.59%/3.88%, 7.32%/4.80%, and 7.27%/4.62% in overall mIoU. These results validate the framework’s effectiveness for urban management, environmental monitoring, and infrastructure planning applications.","PeriodicalId":50269,"journal":{"name":"ISPRS Journal of Photogrammetry and Remote Sensing","volume":"32 1","pages":""},"PeriodicalIF":12.7,"publicationDate":"2025-10-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145229479","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Harmonizing satellite and ground NO2 observations in China: A multi-sensor framework for scenario-specific calibration","authors":"Jianbin Gu ,&nbsp;Xiaoxia Liang ,&nbsp;Shipeng Song ,&nbsp;YiChen Li ,&nbsp;Liangfu Chen ,&nbsp;Jinhua Tao ,&nbsp;Yanfang Tian","doi":"10.1016/j.isprsjprs.2025.09.028","DOIUrl":"10.1016/j.isprsjprs.2025.09.028","url":null,"abstract":"<div><div>Satellite-based nitrogen dioxide (NO₂) retrievals exhibit scenario-dependent discrepancies due to varying environmental conditions, yet systematic evaluations of their performance across heterogeneous regions remain limited. This study presents a comparative analysis of TROPOspheric Monitoring Instrument (TROPOMI) and Ozone Monitoring Instrument (OMI) NO₂ products over China (2019–2023) using 1700+ ground-based monitoring stations through a tripartite framework: (1) evaluation of spatiotemporal consistency with ground observations, (2) analysis of long-term emission trends using deseasonalized data, and (3) scenario-specific validation across seasonal cycles and extreme pollution episodes. Results demonstrate TROPOMI’s superior performance in capturing fine-scale pollution patterns, showing strong correlations with ground measurements in urban areas (e.g. Beijing: R = 0.81) and during extreme events (R = 0.97), while OMI systematically underestimates urban concentrations by 15 % (R = 0.72). Seasonal analysis reveals that TROPOMI and OMI data correlate much more strongly with ground-based measurements under stable winter conditions (R = 0.85 and 0.82, respectively) than in summer, when performance is affected by photochemical processes (R = 0.23 and 0.13, respectively). A unified error model integrating all analytical components is developed to identify the drivers of satellite-ground discrepancies. Applied to the Beijing’s January 2022 episode, the model attributes the observed biases primarily to extreme pollution events (<em>γ</em> = 0.68). Our results emphasize TROPOMI’s superiority for real-time urban air quality management and OMI’s utility for regional trend assessments. This work provides actionable insights for optimizing satellite-ground monitoring systems, supporting targeted emission control strategies under China’s evolving atmospheric policies.</div></div>","PeriodicalId":50269,"journal":{"name":"ISPRS Journal of Photogrammetry and Remote Sensing","volume":"230 ","pages":"Pages 486-494"},"PeriodicalIF":12.2,"publicationDate":"2025-10-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145221831","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Trajectory design for handheld mobile laser scanning in complex natural forests: a simulation approach","authors":"Karel Kuželka","doi":"10.1016/j.isprsjprs.2025.09.025","DOIUrl":"https://doi.org/10.1016/j.isprsjprs.2025.09.025","url":null,"abstract":"This paper presents a simulation study investigating the impact of mobile laser scanning trajectories on the quality of tree stem reconstruction. The analysis is conducted at the level of individual returns, simulating beams emitted from a moving scanner. A real point cloud acquired with the handheld scanning system GeoSLAM Zeb Horizon was used to assess the beam emission pattern, which was then applied to a simulated motion along various trajectory designs.","PeriodicalId":50269,"journal":{"name":"ISPRS Journal of Photogrammetry and Remote Sensing","volume":"6 1","pages":""},"PeriodicalIF":12.7,"publicationDate":"2025-10-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145229480","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Reshaping spatial–temporal pattern of suspended sediment concentration in the Yangtze River mainstem by damming reservoirs based on nearly 40 years Landsat observations","authors":"Farong Chen ,&nbsp;Guangrui Yang ,&nbsp;Yanhui Dai ,&nbsp;Jiale Jin ,&nbsp;Chu Zhao ,&nbsp;Zhishan Ye ,&nbsp;Jiaming Chen ,&nbsp;Xinyi Zhang ,&nbsp;Tao Huang ,&nbsp;Changchun Huang","doi":"10.1016/j.isprsjprs.2025.09.023","DOIUrl":"10.1016/j.isprsjprs.2025.09.023","url":null,"abstract":"<div><div>The proliferation of dams has significantly disrupted sediment flux from land to ocean, reshaping estuarine geomorphology and altering sediment–associated biogeochemical cycles. However, understanding of how dams influence the spatiotemporal distribution of sediment in the Yangtze River remains limited due to sparse observational records. To addresses this gap, this study introduces a novel satellite–derived framework that leverages 36 years of Landsat imagery to quantify patterns of suspended sediment concentration (SSC) along the river. Given the inherent limitations of traditional SSC monitoring, particularly spectral saturation and scalability constraints when quantifying high concentration over large spatial domains, this study calibrated machine learning models using a combination of a public dataset (N=2410) and a field cruise sampling dataset (N=214). The XGBoost–based model achieved robust predictive performance (Public: R²=0.80, MAE=17.19 mg/L, and RMSE=54.13 mg/L; Cruise: R²=0.88, MAE=9.63 mg/L, and RMSE=13.86 mg/L), enabling detailed mapping of SSC spatiotemporal dynamics along the Yangtze River mainstem. Over the study period, SSC exhibited a pronounced decline, decreasing from 767.23±6.51 mg/L in 1986 to 48.14±0.70 mg/L in 2022. The construction of upstream cascading dams shifted high–SSC zones from the upper reach to the middle and lower reaches, while reversing the upstream sediment regime from erosion to deposition, with an average reservoir accumulation of 161.77 Mt yr⁻¹. Among these dams, the Liyuan (LY), Ludila (LDL), Ahai (AH), Jinanqiao (JAQ), and Three Gorges Dam (TGD) exerted the most pronounced influence, with their commissioning closely aligning with marked SSC reductions and abrupt regime shifts. Cascading dams were identified as the dominant drivers of the reshaped sediment distribution, responsible for 32.30 % of the change, exceeding the contributions of upstream soil and water conservation measures (land use transition: 26.12 %, vegetation restoration: 9.13 %) and climate factors (30.44 %). This study quantifies sediment redistribution in the Yangtze River mainstem, elucidates multidecadal SSC responses to dam construction, and provides a transferable framework for sediment–related environmental assessments in ungauged regions.</div></div>","PeriodicalId":50269,"journal":{"name":"ISPRS Journal of Photogrammetry and Remote Sensing","volume":"230 ","pages":"Pages 469-485"},"PeriodicalIF":12.2,"publicationDate":"2025-10-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145221770","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Generalized satellite-derived bathymetry across spatial and temporal domains: a domain-adaptive deep learning approach with multi-source remote sensing data","authors":"Changda Liu ,&nbsp;Huan Xie ,&nbsp;Kuifeng Luan ,&nbsp;Qi Xu ,&nbsp;Yuan Sun ,&nbsp;Min Ji ,&nbsp;Xiaohua Tong","doi":"10.1016/j.isprsjprs.2025.09.021","DOIUrl":"10.1016/j.isprsjprs.2025.09.021","url":null,"abstract":"<div><div>Accurate bathymetric data are crucial for the protection of marine ecosystems and for various human activities. Satellite-derived bathymetry (SDB) based on an empirical approach has been widely employed to estimate shallow water depths. However, the empirical approaches are regression models dependent on specific data, facing challenges in generalizing across different spatial and temporal domains. These challenges can be attributed to domain shifts caused by variations in water quality, substrate, and atmospheric conditions. This limitation hampers their scalability for large-scale or long-term monitoring applications. In this paper, we propose a domain adaptation-based deep learning model for satellite-derived bathymetry (DA-SDB) to address the limitations. Specifically, the DA-SDB model comprises three components: a feature extractor, a bathymetry predictor, and a domain aligner. The feature extractor integrates a pyramid-like block (PLB) and a physical-assisted block (PAB) to improve the data utilization and extract domain-invariant features. The domain aligner mitigates domain shift by aligning the pseudo-inverse Gram matrices. We conducted spatial and temporal transfer experiments in five diverse study areas (Dongsha Atoll, Bimini Island, South Warden Reef, Hadrah Island, and Mubarraz and Bu Tinah Islands). The DA-SDB model demonstrated significant improvements in generalization ability (with the root-mean-square error (RMSE) and mean absolute percentage error (MAPE) reduced by 0.27 <!--> <!-->m and 21.51 %, respectively) and stability (achieving 11 out of 12 best results) over the state-of-the-art methods. Notably, the DA-SDB model maintains robust accuracy across a wide range of depths, as shown by its agreement with reference bathymetry. It employs a top-of-atmosphere (TOA) reflectance dataset and achieves effective results without atmospheric correction. Utilizing this deep learning method, the bathymetric mapping of remote areas and long-term monitoring of critical regions can be conducted rapidly and cost-effectively.</div></div>","PeriodicalId":50269,"journal":{"name":"ISPRS Journal of Photogrammetry and Remote Sensing","volume":"230 ","pages":"Pages 452-468"},"PeriodicalIF":12.2,"publicationDate":"2025-10-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145221830","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"From GPS to AI: A comprehensive review of Unmanned Aerial Vehicle (UAV) localization solutions","authors":"Fahad Lateef,&nbsp;Mohamed Kas,&nbsp;Yassine Ruichek","doi":"10.1016/j.isprsjprs.2025.09.014","DOIUrl":"10.1016/j.isprsjprs.2025.09.014","url":null,"abstract":"<div><div>Unmanned Aerial Vehicle (UAV) technology has undergone significant advances, revolutionizing aerial operations in various sectors by offering adaptability, cost-effectiveness, mobility, and rapid deployment capabilities. Effective UAV navigation depends on several critical factors, with localization being paramount. This study presents a detailed taxonomy of existing UAV localization solutions, which are classified and thoroughly examined in terms of architectural design, technologies employed, data used, applications, performance, and their respective advantages and limitations. This article offers a more comprehensive and up-to-date review of UAV localization solutions and challenges, incorporating solutions from satellite, radio, vision, inertial, lidar, magnetic, acoustic, ultrasonic, and deep learning technologies, exceeding the scope of related surveys. Furthermore, the study explores the essential datasets crucial for UAV localization research, providing detailed information on their specifications and characteristics. By synthesizing all the information, the article highlights existing challenges and potential avenues for future research to advance UAV localization techniques and enhance comprehension of their efficacy across various operational scenarios.</div></div>","PeriodicalId":50269,"journal":{"name":"ISPRS Journal of Photogrammetry and Remote Sensing","volume":"230 ","pages":"Pages 402-451"},"PeriodicalIF":12.2,"publicationDate":"2025-10-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145221828","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"TripleS: Mitigating multi-task learning conflicts for semantic change detection in high-resolution remote sensing imagery","authors":"Xiaoliang Tan ,&nbsp;Guanzhou Chen ,&nbsp;Xiaodong Zhang ,&nbsp;Tong Wang ,&nbsp;Jiaqi Wang ,&nbsp;Kui Wang ,&nbsp;Tingxuan Miao","doi":"10.1016/j.isprsjprs.2025.09.019","DOIUrl":"10.1016/j.isprsjprs.2025.09.019","url":null,"abstract":"<div><div>Periodical earth observation from multi-temporal high spatial resolution remote sensing imagery (RSI) offers valuable insights into the complex dynamics of land surface changes. Semantic change detection (SCD), cooperating with deep learning (DL) architectures, has evolved from binary change detection (BCD) into an effective technique capable of not only identifying change locations but also specifying land-cover and land-use (LCLU) categories. Recent advancements suggest that SCD can be modeled as a multi-task learning (MTL) framework, involving multiple branches for individual subtasks to process dual RSI inputs, and optimized through joint training. However, limitations remain in the inadequate interactions between bi-temporal branches and semantic-change branches, as well as the pervasive gradient conflicts among subtasks within MTL frameworks, which can lead to counterbalanced performances. To address the above limitations, we propose an MTL-oriented SCD model (MOSCD), which mutually enhances bi-temporal features, while ensuring that representations across the subtask branches are coherently correlated. Furthermore, the TripleS framework is designed to enhance the optimization of the MTL framework through counteracting the conflicting subtask objectives, which incorporates three novel schemes: Stepwise multi-task optimization, Selective parameter binding, and Scheduling for dynamically training MTL bindings. Extensive experiments conducted on three full-coverage land-cover SCD datasets, including one public dataset (HRSCD) and two self-constructed datasets (SC-SCD7 and CC-SCD5), demonstrate that the MOSCD enhanced with TripleS outperforms eleven existing SCD methods and three MTL methods by up to 21.17% on SeK metrics. The robust performances over diverse landscapes and transferability on other componentized benchmarks validate that the MOSCD trained with TripleS is a practicable tool for detecting subtle land-cover changes from high spatial resolution RSI data. Codes and the two constructed datasets will be available at <span><span>https://github.com/StephenApX/MTL-TripleS</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":50269,"journal":{"name":"ISPRS Journal of Photogrammetry and Remote Sensing","volume":"230 ","pages":"Pages 374-401"},"PeriodicalIF":12.2,"publicationDate":"2025-09-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145159309","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"ALM-LED: autonomous LiDAR mapping in underground space with Luojia explorer anti-collision drone","authors":"Shangzhe Sun ,&nbsp;Chi Chen ,&nbsp;Bisheng Yang ,&nbsp;Yuhang Xu ,&nbsp;Leyi Zhao ,&nbsp;Yong He ,&nbsp;Ang Jin ,&nbsp;Liuchun Li","doi":"10.1016/j.isprsjprs.2025.09.016","DOIUrl":"10.1016/j.isprsjprs.2025.09.016","url":null,"abstract":"<div><div>With the development of urbanization, underground spaces have become an important part of human life. Accurately surveying and describing the spatial information of underground spaces is of significant importance. However, the complex environment of underground spaces, often characterized by darkness, narrowness, lack of structure, and GNSS-denied conditions, presents tremendous challenges for intelligent information acquisition and analysis in such environments. To address these challenges, we propose ALM-LED, an autonomous LiDAR mapping framework designed for underground environments, which integrates a cost-effective, Luojia explorer anti-collision drone system featuring a lightweight LiDAR sensor and a carbon fiber frame. This framework consists of two main modules: localization and mapping, and planning and control. Localization and mapping integrates LiDAR point cloud data, IMU data, as well as flight control barometer and magnetometer sensor data, enabling robust localization and high-precision mapping in GNSS-denied underground environments. Planning and control constructs a triple constrained cost function for flight trajectory optimization based on smoothness, dynamic feasibility, and collision penalty terms, providing autonomous flight paths for the anti-collision drone system and combining MAVROS, achieving robust control. To validate the proposed system and methods, we conducted experiments in one simulation scenario and two real-world underground scenarios. The experiments demonstrate that ALM-LED achieves average mapping efficiencies exceeding 100 m³/s in simulated environments and 50 m³/s in real-world scenarios when applied to underground spaces. The flight trajectory estimated by the localization and mapping subsystem is nearly identical to the target trajectory. Point cloud maps with volumes of 879 m³, 26313 m³, 22240m³ and 115m³ were generated in four real-world scenarios, with point cloud map accuracies reaching 0.034m, 0.31m, 0.088m and 0.053m, respectively. The experimental results indicate that ALM-LED can achieve efficient and accurate information acquisition in underground spaces, demonstrating high application potential. To support the research community, the key source code for this work is publicly available at the following repository: <span><span>https://github.com/DCSI2022/ALM-LED</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":50269,"journal":{"name":"ISPRS Journal of Photogrammetry and Remote Sensing","volume":"230 ","pages":"Pages 346-373"},"PeriodicalIF":12.2,"publicationDate":"2025-09-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145159310","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Reconstructing wildland fire burned area for Asian Russia (1979–2000) using AVHRR GAC satellite data to provide an improved baseline for assessing long-term change","authors":"Donald R. Cahoon Jr. ,&nbsp;Amber J. Soja ,&nbsp;Brian J. Stocks ,&nbsp;Stefano Potter ,&nbsp;Natasha Jurko ,&nbsp;Emily Gargulinski ,&nbsp;Brendan M. Rogers ,&nbsp;Susan G. Conard","doi":"10.1016/j.isprsjprs.2025.09.011","DOIUrl":"10.1016/j.isprsjprs.2025.09.011","url":null,"abstract":"<div><div>Wildland fire is a vital ecological disturbance at northern latitudes that has strong interactions with weather and climate systems. Multi-decadal fire records are critical for assessing changing fire regimes and vegetation mosaic patterns. While such records are available for Alaska, Canada, and Fennoscandia, accurate pre-2000 data for Russia are notably lacking. Continuous moderate-to-high-resolution data for Asian Russia are not available before the 2000s. In this study we defined fire scars using Advanced Very High Resolution Radiometer (AVHRR) Global Area Coverage (GAC) data to develop a continuous, long-term, burned-area database for Asian Russia that spans from 1979 through 2000. We generated monthly composites of fire scars from daily GAC data and used a combination of aerosol and visible smoke data to confirm that observed spectral changes were due to fire and to determine dates of active burning. Accuracy of burned areas was evaluated using available Landsat Thematic Mapper (TM) data and correlations with previously published burn scar data. The coefficient of determination (R²) for linear regressions between Landsat validation burned areas and GAC data was 0.84 for all fires sampled and 0.97 for large fires greater than 10,000 ha (ha), north of 53 degrees latitude. Omission-Commission analysis also show higher accuracy with larger fires. The overall comparison with previously published large burn-scar data had an R² of 0.88. The largest errors were with fires less than 10,000 ha, which make up less than 7 % of the burned area. We present seasonal fire patterns and spatial and ecozone distribution of burned areas. In a typical season, fires started in the south and spread to the north over spring and early summer. We observed high interannual variation in the spatial patterns of burned area. Total annual burned areas ranged from 0.4 to 11.9 million hectares (Mha), with an average burned area of 4.8 Mha per year. Our estimates for most years are several times higher than official Russian burned-area reports and are typically larger than burned-area data reported in previous publications. Our data represent the first validated long-term historical burned-area data for Asian Russia, which provides an essential basis for analyses of the interactions between these diverse and unique ecoregions, fire regimes, and weather and climate feedbacks. When combined with existing data from other northern regions, our data will enable accurate assessments of long-term fire patterns and fire-climate interactions across the critical boreal-Arctic region for almost 50 years.</div></div>","PeriodicalId":50269,"journal":{"name":"ISPRS Journal of Photogrammetry and Remote Sensing","volume":"230 ","pages":"Pages 318-345"},"PeriodicalIF":12.2,"publicationDate":"2025-09-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145159311","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"AdaptVFMs-RSCD: Advancing Remote Sensing Change Detection from binary to semantic with SAM and CLIP","authors":"Wandong Jiang,&nbsp;Yuli Sun,&nbsp;Lin Lei,&nbsp;Gangyao Kuang,&nbsp;Kefeng Ji","doi":"10.1016/j.isprsjprs.2025.09.010","DOIUrl":"10.1016/j.isprsjprs.2025.09.010","url":null,"abstract":"<div><div>Remote Sensing Change Detection (RSCD) is essential for identifying surface changes from remote sensing images (RSIs) and plays a crucial role in land-use planning and disaster assessment. Despite advancements in RSI resolution and AI, most RSCD datasets are binary, hindering the transition to semantic change detection. Vision Foundation Models (VFMs), such as the Segment Anything Model (SAM), introduce new possibilities with robust zero-shot semantic segmentation capabilities, but face challenges with RSIs due to their unique characteristics, such as diverse perspectives and scale variations. To address these challenges, an enhanced RSCD method, AdaptVFMs-RSCD, has been proposed. This method integrates SAM with Contrastive Language-Image Pre-training (CLIP), capitalizing on CLIP’s ability to establish broad correspondences between images and text and to classify unseen image categories. This integration markedly improves the recognition of land cover types, better aligning VFMs with the specific requirements of RSCD. Additionally, a remote sensing VFM fine-tuning dataset was also developed to further enhance SAM’s segmentation performance on RSIs. Furthermore, a semantic information-based change detection module was designed to fully leverage both the change information and semantic information provided by VFMs, achieving state-of-the-art F1 and mIoU scores on the DSIFN (66.94%, 67.84%), CLCD (76.12%, 78.80%), and SYSU (81.14%, 78.35%) datasets. Notably, the comprehensive metrics F1 and mIoU on the SYSU dataset exceeded the second-best scores by 17.57% and 17.77%, respectively. AdaptVFMs-RSCD also facilitates the conversion of binary change detection datasets into semantic change detection datasets, advancing semantic change detection and expanding the application of vision language models and VFMs in remote sensing.</div></div>","PeriodicalId":50269,"journal":{"name":"ISPRS Journal of Photogrammetry and Remote Sensing","volume":"230 ","pages":"Pages 304-317"},"PeriodicalIF":12.2,"publicationDate":"2025-09-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145159312","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}