International journal of applied earth observation and geoinformation : ITC journal最新文献_第2页

Salinity-induced global pattern of atmospheric water constraints on mangrove photosynthetic activity revealed by time series Sentinel-2 data Sentinel-2时间序列数据揭示的盐度诱导的全球大气水分约束红树林光合活性格局

IF 8.6 Q1 REMOTE SENSING

International journal of applied earth observation and geoinformation : ITC journal

Pub Date : 2026-03-01 Epub Date: 2026-02-17 DOI: 10.1016/j.jag.2026.105170

Yanjie Liu , Yueting Deng , Hui Luo , Nengwang Chen , Yougan Chen , Zhenong Jin , Xu Wang , Hongsheng Zhang , Xudong Zhu

Atmospheric drought stress limits mangrove photosynthetic activity, and this constraint can be further amplified by high salinity, yet their combined global effects remain poorly understood. Here, we integrated multi-source Earth observation and geoinformation datasets, including Sentinel-2 red-edge position (a proxy for canopy photosynthetic activity), vapor pressure deficit from TerraClimate, seawater salinity from Copernicus reanalysis, to investigate how salinity regulates the sensitivity of mangrove photosynthesis to atmospheric drought stress during 2019–2023. Datasets were harmonized and analyzed through reproducible geoinformation workflows at 10 m–0.5° resolutions, enabling large-scale coupling analyses between remote sensing proxies and climate drivers. We found that drought stress constrained mangrove photosynthetic activity worldwide, with stronger limitations in tropical savannahs than in tropical rainforests. Marine mangroves exposed to persistent high salinity were more sensitive than estuarine mangroves influenced by freshwater inflow. These results reveal a global pattern in which salinity amplifies atmospheric water constraints on mangrove photosynthesis. Mangroves in dry climates and high-salinity habitats are therefore most vulnerable to future warming and drying. Our findings confirm that integrating multi-source satellite observations with geoinformation analysis provides an effective, large-scale approach for assessing vegetation vulnerability and identifying conservation priorities in climate-sensitive mangrove ecosystems.

大气干旱胁迫限制了红树林的光合作用，而这种限制可能会被高盐度进一步放大，但它们的综合全球影响仍然知之甚少。在此，我们整合了多源地球观测和地理信息数据集，包括Sentinel-2红边位置（冠层光合活动的代理）、terrclimate的蒸汽压亏缺、哥白尼再分析的海水盐度，研究盐度如何调节红树林光合作用对大气干旱胁迫的敏感性。通过10 m-0.5°分辨率的可重复地理信息工作流对数据集进行协调和分析，实现了遥感代理与气候驱动因素之间的大规模耦合分析。我们发现，干旱胁迫在全球范围内限制了红树林的光合活动，热带稀树草原的限制比热带雨林更强。持续高盐度环境下的海洋红树林比河口红树林对淡水流入的影响更为敏感。这些结果揭示了一个全球模式，其中盐度放大了大气水分对红树林光合作用的限制。因此，干旱气候和高盐度栖息地的红树林最容易受到未来变暖和干燥的影响。我们的研究结果证实，将多源卫星观测与地理信息分析相结合，为评估气候敏感红树林生态系统的植被脆弱性和确定保护重点提供了一种有效的大规模方法。

{"title":"Salinity-induced global pattern of atmospheric water constraints on mangrove photosynthetic activity revealed by time series Sentinel-2 data","authors":"Yanjie Liu , Yueting Deng , Hui Luo , Nengwang Chen , Yougan Chen , Zhenong Jin , Xu Wang , Hongsheng Zhang , Xudong Zhu","doi":"10.1016/j.jag.2026.105170","DOIUrl":"10.1016/j.jag.2026.105170","url":null,"abstract":"<div><div>Atmospheric drought stress limits mangrove photosynthetic activity, and this constraint can be further amplified by high salinity, yet their combined global effects remain poorly understood. Here, we integrated multi-source Earth observation and geoinformation datasets, including Sentinel-2 red-edge position (a proxy for canopy photosynthetic activity), vapor pressure deficit from TerraClimate, seawater salinity from Copernicus reanalysis, to investigate how salinity regulates the sensitivity of mangrove photosynthesis to atmospheric drought stress during 2019–2023. Datasets were harmonized and analyzed through reproducible geoinformation workflows at 10 m–0.5° resolutions, enabling large-scale coupling analyses between remote sensing proxies and climate drivers. We found that drought stress constrained mangrove photosynthetic activity worldwide, with stronger limitations in tropical savannahs than in tropical rainforests. Marine mangroves exposed to persistent high salinity were more sensitive than estuarine mangroves influenced by freshwater inflow. These results reveal a global pattern in which salinity amplifies atmospheric water constraints on mangrove photosynthesis. Mangroves in dry climates and high-salinity habitats are therefore most vulnerable to future warming and drying. Our findings confirm that integrating multi-source satellite observations with geoinformation analysis provides an effective, large-scale approach for assessing vegetation vulnerability and identifying conservation priorities in climate-sensitive mangrove ecosystems.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"147 ","pages":"Article 105170"},"PeriodicalIF":8.6,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146209250","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Explainable artificial intelligence reveals heterogeneous erosion responses to extreme rainfall: A new framework for conservation prioritization 可解释的人工智能揭示了极端降雨对异质性侵蚀的响应：保护优先级的新框架

IF 8.6 Q1 REMOTE SENSING

International journal of applied earth observation and geoinformation : ITC journal

Pub Date : 2026-03-01 Epub Date: 2026-02-21 DOI: 10.1016/j.jag.2026.105181

Dongling Ma , Shuangyun Peng , Zhiqiang Lin , Jiaying Zhu , Xianchun Pan , Yuanmei Jiao , Ziyi Zhu , Shuangfu Shi , Biting Cui , Rong Jin

Effectively combating soil erosion under intensifying extreme precipitation requires moving beyond broad-scale assessments to precision-targeted interventions. Current approaches often fail to capture the differentiated erosion responses across diverse landscapes, limiting the efficacy of conservation investments. Here, we introduce a novel framework that combines modeling (RUSLE) with Explainable artificial intelligence (LightGBM model with SHAP method) and multi-criteria decision analysis to prioritize soil and water conservation (SWC) efforts. Applied to 20 SWC zones in the ecologically critical Yunnan Province, China, our analysis reveals that maximum 5-day precipitation (RX5) is the paramount driver of erosion, superseding other precipitation metrics. We uncover distinct regional response mechanisms: erosion in northwestern cold alpine zones is governed by cumulative rainfall, while erosion in temperate and tropical zones is triggered by short-duration, high-intensity events. This mechanistic understanding enabled the robust identification of the Northwest Alpine Gorge and Western Broad Valley zones as highest-priority areas demanding urgent action. By systematically diagnosing the primary climatic drivers and identifying the most vulnerable regions, our framework provides a powerful, replicable blueprint for optimizing conservation resources and enhancing climate resilience in mountainous ecosystems globally.

在极端降水加剧的情况下，有效防治土壤侵蚀需要从大规模评估转向有针对性的精确干预。目前的方法往往不能捕捉不同景观的不同侵蚀响应，限制了保护投资的有效性。在这里，我们引入了一个新的框架，将建模（RUSLE）与可解释的人工智能（LightGBM模型与SHAP方法）和多标准决策分析相结合，以优先考虑水土保持（SWC）工作。应用于生态临界型云南省的20个SWC带，我们的分析表明，最大5天降水（RX5）是侵蚀的主要驱动因素，取代了其他降水指标。我们发现了不同的区域响应机制：西北冷高寒地区的侵蚀受累积降雨控制，而温带和热带地区的侵蚀由短时间、高强度的事件触发。这种机制的理解使西北阿尔卑斯峡谷和西部宽谷地区成为需要紧急行动的最优先区域。通过系统地诊断主要气候驱动因素并确定最脆弱的地区，我们的框架为优化保护资源和增强全球山区生态系统的气候适应能力提供了一个强大的、可复制的蓝图。

{"title":"Explainable artificial intelligence reveals heterogeneous erosion responses to extreme rainfall: A new framework for conservation prioritization","authors":"Dongling Ma , Shuangyun Peng , Zhiqiang Lin , Jiaying Zhu , Xianchun Pan , Yuanmei Jiao , Ziyi Zhu , Shuangfu Shi , Biting Cui , Rong Jin","doi":"10.1016/j.jag.2026.105181","DOIUrl":"10.1016/j.jag.2026.105181","url":null,"abstract":"<div><div>Effectively combating soil erosion under intensifying extreme precipitation requires moving beyond broad-scale assessments to precision-targeted interventions. Current approaches often fail to capture the differentiated erosion responses across diverse landscapes, limiting the efficacy of conservation investments. Here, we introduce a novel framework that combines modeling (RUSLE) with Explainable artificial intelligence (LightGBM model with SHAP method) and multi-criteria decision analysis to prioritize soil and water conservation (SWC) efforts. Applied to 20 SWC zones in the ecologically critical Yunnan Province, China, our analysis reveals that maximum 5-day precipitation (RX5) is the paramount driver of erosion, superseding other precipitation metrics. We uncover distinct regional response mechanisms: erosion in northwestern cold alpine zones is governed by cumulative rainfall, while erosion in temperate and tropical zones is triggered by short-duration, high-intensity events. This mechanistic understanding enabled the robust identification of the Northwest Alpine Gorge and Western Broad Valley zones as highest-priority areas demanding urgent action. By systematically diagnosing the primary climatic drivers and identifying the most vulnerable regions, our framework provides a powerful, replicable blueprint for optimizing conservation resources and enhancing climate resilience in mountainous ecosystems globally.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"147 ","pages":"Article 105181"},"PeriodicalIF":8.6,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147278340","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Detection and accuracy of a geomorphic Proxy-Based shoreline indicator in PlanetScope imagery PlanetScope图像中基于地貌代理的海岸线指标的检测与精度

IF 8.6 Q1 REMOTE SENSING

International journal of applied earth observation and geoinformation : ITC journal

Pub Date : 2026-03-01 Epub Date: 2026-02-20 DOI: 10.1016/j.jag.2026.105173

Joshua T. Kelly, Suvam Patel

Satellite-derived shoreline mapping is a common technique for quantifying geomorphic shoreline change. However, shoreline positions derived using moderate-resolution data have questionable accuracy, are influenced by metocean conditions, and are typically based on the boundary of a binarized spectral index rather than a visible geomorphic indicator, such as the high-water line (HWL). PlanetScope (PS) can visualize the location of the HWL by detecting the spectral reflectance differences between wet and dry sediment along a sandy beach surface due to its improved spatial resolution of 3 m/pixel. The accuracy of nine HWL proxy shoreline positions is assessed by comparison to a contemporaneous mean high water (MHW) shoreline delineated across Moro Beach, CA, using a digital elevation model created from RTK GPS-corrected Unmanned Aircraft System imagery. The offset between the PS-derived HWL and UAS-derived MHW positions (Δd) was measured every 10 m in the alongshore direction using the Digital Shoreline Analysis System for each HWL dataset. A significant exponential relationship was observed between Δd and the tide height at the time of PS image acquisition, whereby the HWL shoreline was located further landward (seaward) during higher (lower) tides. The HWL shoreline, when acquired at or near low tide, was spatially coincident with the MHW shoreline, the most reliable yet cost-prohibitive shoreline proxy. PlanetScope’s advancement in spatiotemporal resolution introduces a new approach to satellite-derived shoreline mapping, one that is based on a geomorphic proxy position and a minimized influence of tide heights.

卫星岸线制图是量化地貌岸线变化的常用技术。然而，使用中等分辨率数据得出的海岸线位置精度有问题，受海洋气象条件的影响，并且通常基于二值化光谱指数的边界，而不是可见的地貌指标，如高潮线（HWL）。PlanetScope （PS）提高了3米/像素的空间分辨率，通过检测沙滩表面湿沉积物和干沉积物的光谱反射率差异，可以可视化HWL的位置。通过使用RTK gps校正无人机系统图像创建的数字高程模型，将9个HWL代理海岸线位置的准确性与横跨CA Moro Beach的同期平均高水位（MHW）海岸线进行比较。利用数字海岸线分析系统对每个HWL数据集在岸线方向每10米测量一次ps衍生HWL和uas衍生MHW位置之间的偏移量（Δd）。在PS图像采集时，Δd与潮高之间存在显著的指数关系，即在高（低）潮时，HWL海岸线更靠近陆地（向海）。当在退潮时或接近退潮时，HWL海岸线在空间上与MHW海岸线一致，MHW海岸线是最可靠但成本过高的海岸线代理。PlanetScope在时空分辨率方面的进步为卫星衍生的海岸线测绘引入了一种新的方法，这种方法基于地貌代理位置和最小化潮汐高度的影响。

{"title":"Detection and accuracy of a geomorphic Proxy-Based shoreline indicator in PlanetScope imagery","authors":"Joshua T. Kelly, Suvam Patel","doi":"10.1016/j.jag.2026.105173","DOIUrl":"10.1016/j.jag.2026.105173","url":null,"abstract":"<div><div>Satellite-derived shoreline mapping is a common technique for quantifying geomorphic shoreline change. However, shoreline positions derived using moderate-resolution data have questionable accuracy, are influenced by metocean conditions, and are typically based on the boundary of a binarized spectral index rather than a visible geomorphic indicator, such as the high-water line (HWL). PlanetScope (PS) can visualize the location of the HWL by detecting the spectral reflectance differences between wet and dry sediment along a sandy beach surface due to its improved spatial resolution of 3 m/pixel. The accuracy of nine HWL proxy shoreline positions is assessed by comparison to a contemporaneous mean high water (MHW) shoreline delineated across Moro Beach, CA, using a digital elevation model created from RTK GPS-corrected Unmanned Aircraft System imagery. The offset between the PS-derived HWL and UAS-derived MHW positions (Δd) was measured every 10 m in the alongshore direction using the Digital Shoreline Analysis System for each HWL dataset. A significant exponential relationship was observed between Δd and the tide height at the time of PS image acquisition, whereby the HWL shoreline was located further landward (seaward) during higher (lower) tides. The HWL shoreline, when acquired at or near low tide, was spatially coincident with the MHW shoreline, the most reliable yet cost-prohibitive shoreline proxy. PlanetScope’s advancement in spatiotemporal resolution introduces a new approach to satellite-derived shoreline mapping, one that is based on a geomorphic proxy position and a minimized influence of tide heights.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"147 ","pages":"Article 105173"},"PeriodicalIF":8.6,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147278288","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Evaluating superpixel algorithms for standing dead tree delineation using aerial orthoimagery 利用航空正射影成像评估直立枯树描绘的超像素算法

IF 8.6 Q1 REMOTE SENSING

International journal of applied earth observation and geoinformation : ITC journal

Pub Date : 2026-03-01 Epub Date: 2026-02-17 DOI: 10.1016/j.jag.2026.105180

Igor Pawelec, Paweł Hawryło, Paweł Netzel, Jarosław Socha

High-resolution remote sensing data are essential for monitoring forest health and detecting changes such as tree mortality. This study evaluates low-level segmentation methods for delineating standing dead trees (SDTs) using widely available true-color (RGB) aerial orthoimagery, independent of near-infrared (NIR) or LiDAR data. The analysis was conducted across eight forest sites in northern Poland, dominated by coniferous species such as Scots pine (Pinus sylvestris L.) and Norway spruce (Picea abies L.).

Four representative superpixel-based algorithms were tested — Simple Linear Iterative Clustering (SLIC), its zero-parameter variant (SLIC0), scale-adaptive superpixels (adaptels), and a spatially regularized watershed transform (waterpixels). All methods represent preprocessing approaches designed to reduce image complexity and preserve meaningful spectral–spatial structures prior to object-based image analysis (OBIA). In addition, the impact of converting imagery to the perceptually uniform CIELAB color space was assessed to enhance spectral separability and reduce illumination effects. Segmentation accuracy was evaluated against a manually verified reference dataset of 1,200 SDT crowns using multiple quality metrics.

The results indicate that the adaptels algorithm, particularly when combined with the CIELAB color transformation, achieved the most balanced performance across all evaluation metrics, defined by a simultaneous reduction of segmentation fragmentation and boundary generalization errors while maintaining high overall detection accuracy. This combination proved to be an efficient and cost-effective solution for SDT segmentation using standard RGB orthophotos. The findings highlight the potential of perceptually uniform color transformations as practical tools for scalable, reproducible, and low-cost forest monitoring. The study also provides a reference database of standing dead trees to support further research and future integration with deep learning-based detection frameworks.

高分辨率遥感数据对于监测森林健康和探测树木死亡率等变化至关重要。本研究评估了使用广泛可用的真彩色（RGB）航空正射成像，独立于近红外（NIR）或激光雷达数据，描绘直立死树（SDTs）的低水平分割方法。该分析在波兰北部的8个森林地点进行，主要是针叶物种，如苏格兰松（Pinus sylvestris L.）和挪威云杉（Picea abies L.）。

{"title":"Evaluating superpixel algorithms for standing dead tree delineation using aerial orthoimagery","authors":"Igor Pawelec, Paweł Hawryło, Paweł Netzel, Jarosław Socha","doi":"10.1016/j.jag.2026.105180","DOIUrl":"10.1016/j.jag.2026.105180","url":null,"abstract":"<div><div>High-resolution remote sensing data are essential for monitoring forest health and detecting changes such as tree mortality. This study evaluates low-level segmentation methods for delineating standing dead trees (SDTs) using widely available true-color (RGB) aerial orthoimagery, independent of near-infrared (NIR) or LiDAR data. The analysis was conducted across eight forest sites in northern Poland, dominated by coniferous species such as Scots pine (<em>Pinus sylvestris</em> L.) and Norway spruce (<em>Picea abies</em> L.).</div><div>Four representative superpixel-based algorithms were tested — Simple Linear Iterative Clustering (SLIC), its zero-parameter variant (SLIC0), scale-adaptive superpixels (adaptels), and a spatially regularized watershed transform (waterpixels). All methods represent preprocessing approaches designed to reduce image complexity and preserve meaningful spectral–spatial structures prior to object-based image analysis (OBIA). In addition, the impact of converting imagery to the perceptually uniform CIELAB color space was assessed to enhance spectral separability and reduce illumination effects. Segmentation accuracy was evaluated against a manually verified reference dataset of 1,200 SDT crowns using multiple quality metrics.</div><div>The results indicate that the adaptels algorithm, particularly when combined with the CIELAB color transformation, achieved the most balanced performance across all evaluation metrics, defined by a simultaneous reduction of segmentation fragmentation and boundary generalization errors while maintaining high overall detection accuracy. This combination proved to be an efficient and cost-effective solution for SDT segmentation using standard RGB orthophotos. The findings highlight the potential of perceptually uniform color transformations as practical tools for scalable, reproducible, and low-cost forest monitoring. The study also provides a reference database of standing dead trees to support further research and future integration with deep learning-based detection frameworks.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"147 ","pages":"Article 105180"},"PeriodicalIF":8.6,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147278290","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Spectral-Feature-Driven photovoltaic Detection: A universal Physics-Based index for rapid Localization 光谱特征驱动的光伏检测：一种通用的基于物理的快速定位指标

IF 8.6 Q1 REMOTE SENSING

International journal of applied earth observation and geoinformation : ITC journal

Pub Date : 2026-03-01 Epub Date: 2026-02-28 DOI: 10.1016/j.jag.2026.105164

Shuang He , Qingjiu Tian , Jia Tian , Lina Hao

Photovoltaic (PV) energy is critical to the transition towards a net-zero economy and plays a vital role in meeting the Sustainable Development Goals (SDGs), particularly regarding affordable clean energy (SDG 7) and climate action (SDG 13). Timely and accurate acquisition of the spatial distribution of PV installations is critical for regional energy planning, capacity estimation, and policy adjustment. However, accurately detecting PV installations remains challenging due to their environmental complexity and structural diversity. Through multi-platform spectral analysis (including Sentinel-2, Landsat-8, and GF-2 imagery), this study identifies distinctive spectral reflectance properties of PV materials, characterized by a prominent peak in the 400–500 nm range and significantly lower reflectance in the visible to near-infrared spectrum compared to natural landscapes, while exhibiting higher reflectance than water bodies. Leveraging physics-based spectral signatures that remain consistent across diverse geographical settings, we introduce the Spectral Ratio-Normalized Difference Solar Photovoltaic Panel Index (SPPI), a universal approach for efficient PV detection using optical satellite imagery. Quantitative validation across multiple regions (urban, rural, and mountainous environments) demonstrates that SPPI achieves exceptional performance with 94.34% overall accuracy and a robust Kappa coefficient of 0.778, outperforming existing index-based methodologies while producing results comparable to more computationally intensive deep learning approaches. The SPPI methodology’s distinctive advantage lies in its ability to generate precise PV polygon boundaries while maintaining computational efficiency, enabling rapid large-scale mapping without specialized hardware requirements. While installation variations and extreme viewing angles may affect performance, the physics-based nature of the index ensures consistent results under normal imaging conditions. This universal, computationally efficient approach facilitates effective PV installation monitoring and energy capacity estimation, enhancing renewable energy analytics for carbon neutrality initiatives.

光伏（PV）能源对于向净零经济过渡至关重要，在实现可持续发展目标（SDG）方面发挥着至关重要的作用，特别是在负担得起的清洁能源（SDG 7）和气候行动（SDG 13）方面。及时准确地获取光伏装置的空间分布对于区域能源规划、容量估计和政策调整至关重要。然而，由于其环境的复杂性和结构的多样性，准确检测光伏装置仍然具有挑战性。通过多平台的光谱分析（包括Sentinel-2、Landsat-8和GF-2图像），本研究发现了光伏材料独特的光谱反射率特性，其特征是在400-500 nm范围内有一个突出的峰，与自然景观相比，可见光到近红外光谱的反射率明显较低，而反射率高于水体。利用在不同地理环境中保持一致的基于物理的光谱特征，我们介绍了光谱比率归一化差分太阳能光伏板指数（SPPI），这是一种利用光学卫星图像进行高效光伏检测的通用方法。跨多个地区（城市、农村和山区环境）的定量验证表明，SPPI实现了卓越的性能，总体准确率为94.34%，Kappa系数为0.778，优于现有的基于指数的方法，同时产生的结果可与计算密集型的深度学习方法相媲美。SPPI方法的独特优势在于它能够生成精确的PV多边形边界，同时保持计算效率，无需专门的硬件要求即可实现快速大规模映射。虽然安装变化和极端视角可能会影响性能，但该指数基于物理的性质确保了在正常成像条件下的一致结果。这种通用的、计算效率高的方法促进了有效的光伏安装监测和能源容量估计，增强了碳中和倡议的可再生能源分析。

{"title":"Spectral-Feature-Driven photovoltaic Detection: A universal Physics-Based index for rapid Localization","authors":"Shuang He , Qingjiu Tian , Jia Tian , Lina Hao","doi":"10.1016/j.jag.2026.105164","DOIUrl":"10.1016/j.jag.2026.105164","url":null,"abstract":"<div><div>Photovoltaic (PV) energy is critical to the transition towards a net-zero economy and plays a vital role in meeting the Sustainable Development Goals (SDGs), particularly regarding affordable clean energy (SDG 7) and climate action (SDG 13). Timely and accurate acquisition of the spatial distribution of PV installations is critical for regional energy planning, capacity estimation, and policy adjustment. However, accurately detecting PV installations remains challenging due to their environmental complexity and structural diversity. Through multi-platform spectral analysis (including Sentinel-2, Landsat-8, and GF-2 imagery), this study identifies distinctive spectral reflectance properties of PV materials, characterized by a prominent peak in the 400–500 nm range and significantly lower reflectance in the visible to near-infrared spectrum compared to natural landscapes, while exhibiting higher reflectance than water bodies. Leveraging physics-based spectral signatures that remain consistent across diverse geographical settings, we introduce the Spectral Ratio-Normalized Difference Solar Photovoltaic Panel Index (SPPI), a universal approach for efficient PV detection using optical satellite imagery. Quantitative validation across multiple regions (urban, rural, and mountainous environments) demonstrates that SPPI achieves exceptional performance with 94.34% overall accuracy and a robust Kappa coefficient of 0.778, outperforming existing index-based methodologies while producing results comparable to more computationally intensive deep learning approaches. The SPPI methodology’s distinctive advantage lies in its ability to generate precise PV polygon boundaries while maintaining computational efficiency, enabling rapid large-scale mapping without specialized hardware requirements. While installation variations and extreme viewing angles may affect performance, the physics-based nature of the index ensures consistent results under normal imaging conditions. This universal, computationally efficient approach facilitates effective PV installation monitoring and energy capacity estimation, enhancing renewable energy analytics for carbon neutrality initiatives.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"147 ","pages":"Article 105164"},"PeriodicalIF":8.6,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147360590","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

SpatialLLM: From multi-modality data to urban spatial intelligence SpatialLLM：从多模态数据到城市空间智能

IF 8.6 Q1 REMOTE SENSING

International journal of applied earth observation and geoinformation : ITC journal

Pub Date : 2026-03-01 Epub Date: 2026-02-25 DOI: 10.1016/j.jag.2026.105177

Jiabin Chen , Haiping Wang , Jinpeng Li , Yuan Liu , Zhen Dong , Bisheng Yang

We propose SpatialLLM, an integrated framework advancing spatial intelligence tasks in complex urban scenes. Unlike previous methods requiring specialized geographic analysis tools or domain expertise, SpatialLLM leverages the inherent reasoning capabilities of pre-trained Large Language Models (LLMs) to address various spatial intelligence tasks. The core of SpatialLLM lies in constructing detailed and structured scene descriptions from raw spatial data to prompt LLMs for scene-based analysis. Extensive experiments demonstrate that, with our designs, general-purpose LLMs can accurately perceive spatial distribution information and execute advanced spatial intelligence tasks, including urban planning, ecological analysis, traffic management. We further investigate key factors influencing LLM performance in urban analysis, such as multi-field knowledge, context length, and reasoning ability. We hope that SpatialLLM offers a viable perspective for intelligent urban analysis and management. The code and dataset are available at https://github.com/WHU-USI3DV/SpatialLLM.

我们提出了SpatialLLM，这是一个在复杂城市场景中推进空间智能任务的集成框架。与以往需要专门地理分析工具或领域专业知识的方法不同，SpatialLLM利用预先训练的大型语言模型（llm）的固有推理能力来解决各种空间智能任务。SpatialLLM的核心在于从原始空间数据中构建详细的、结构化的场景描述，以提示llm进行基于场景的分析。大量的实验表明，通过我们的设计，通用llm可以准确地感知空间分布信息并执行高级空间智能任务，包括城市规划，生态分析，交通管理。我们进一步研究了影响法学硕士在城市分析中表现的关键因素，如多领域知识、语境长度和推理能力。我们希望SpatialLLM为智慧城市分析和管理提供一个可行的视角。代码和数据集可从https://github.com/WHU-USI3DV/SpatialLLM获得。

{"title":"SpatialLLM: From multi-modality data to urban spatial intelligence","authors":"Jiabin Chen , Haiping Wang , Jinpeng Li , Yuan Liu , Zhen Dong , Bisheng Yang","doi":"10.1016/j.jag.2026.105177","DOIUrl":"10.1016/j.jag.2026.105177","url":null,"abstract":"<div><div>We propose SpatialLLM, an integrated framework advancing spatial intelligence tasks in complex urban scenes. Unlike previous methods requiring specialized geographic analysis tools or domain expertise, SpatialLLM leverages the inherent reasoning capabilities of pre-trained Large Language Models (LLMs) to address various spatial intelligence tasks. The core of SpatialLLM lies in constructing detailed and structured scene descriptions from raw spatial data to prompt LLMs for scene-based analysis. Extensive experiments demonstrate that, with our designs, general-purpose LLMs can accurately perceive spatial distribution information and execute advanced spatial intelligence tasks, including urban planning, ecological analysis, traffic management. We further investigate key factors influencing LLM performance in urban analysis, such as multi-field knowledge, context length, and reasoning ability. We hope that SpatialLLM offers a viable perspective for intelligent urban analysis and management. The code and dataset are available at <span><span>https://github.com/WHU-USI3DV/SpatialLLM</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"147 ","pages":"Article 105177"},"PeriodicalIF":8.6,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147360639","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Mountains of error: UAV LiDAR benchmarking of GEDI and GEDI-derived global canopy height products in steep Himalayan forests 误差山：在陡峭的喜马拉雅森林中GEDI和GEDI衍生的全球冠层高度产品的无人机激光雷达基准测试

IF 8.6 Q1 REMOTE SENSING

International journal of applied earth observation and geoinformation : ITC journal

Pub Date : 2026-03-01 Epub Date: 2026-02-24 DOI: 10.1016/j.jag.2026.105203

Niti B. Mishra , Paras Bikram Singh

Accurate canopy-height information in steep mountain forests is essential for biomass estimation, habitat characterization, and monitoring climate-sensitive ecotones, yet the reliability of spaceborne LiDAR height products in rugged terrain remains poorly constrained. Here we benchmarked Global Ecosystem Dynamics Investigation (GEDI) footprint canopy-height metrics and two GEDI-derived global canopy-height products (GFCH 2019 and GCH 2020) using a 1 m UAV-LiDAR canopy height model acquired along a ∼ 2,000 m elevational transect in the central Himalayas, Nepal. We show that GEDI RH95, paired with the UAV P95 reference in a like-with-like percentile comparison, reproduces the overall canopy-height distribution with minimal systematic bias but substantial footprint-scale dispersion, with errors that are spatially structured by terrain and sub-footprint heterogeneity. A footprint-shift experiment indicates that plausible horizontal misregistration contributes comparatively little to overall uncertainty reinforcing that topography and within-footprint variability dominate the error budget in this setting. At the map scale, GFCH 2019 captures broad elevational patterns but systematically underestimates tall canopy heights and exhibits terrain-dependent residual structure, whereas GCH 2020 shows widespread positive bias and spatially patchy coverage over the LiDAR-mapped domain. Together, these results provide a rare UAV-LiDAR benchmark for interpreting GEDI footprint heights and GEDI-derived canopy height models in a data-scarce, high-relief mountain forest, and highlight the need for terrain-aware uncertainty characterization and local calibration when these products are used for biomass, habitat, and treeline/upper-montane applications.

在陡峭的山地森林中，准确的冠层高度信息对于生物量估算、栖息地表征和监测气候敏感过渡带至关重要，但在崎岖地形中，星载激光雷达高度产品的可靠性仍然受到很好的限制。在这里，我们使用沿尼泊尔喜马拉雅中部海拔约2000米的样带获取的1米无人机-激光雷达冠层高度模型，对全球生态系统动力学调查（GEDI）足迹冠层高度指标和两个GEDI衍生的全球冠层高度产品（GFCH 2019和GCH 2020）进行基准测试。研究表明，GEDI RH95与无人机P95参考数据进行了相似的百分位数比较，以最小的系统偏差再现了总体冠层高度分布，但存在大量的足迹尺度离散，其误差在空间上由地形和子足迹异质性构成。一项足迹移动实验表明，合理的水平错配对总体不确定性的贡献相对较小，这进一步表明，地形和足迹内的可变性在这种情况下主导了误差预算。在地图尺度上，GFCH 2019捕获了广泛的海拔格局，但系统性地低估了高冠层高度，并显示出地形相关的残余结构，而GCH 2020在激光雷达绘制的区域内显示出广泛的正偏差和空间斑块覆盖。总之，这些结果为在数据稀缺的高山森林中解释GEDI足迹高度和GEDI衍生的冠层高度模型提供了一个罕见的无人机- lidar基准，并强调了当这些产品用于生物量、栖息地和林木线/高山应用时，地形感知不确定性表征和局部校准的必要性。

{"title":"Mountains of error: UAV LiDAR benchmarking of GEDI and GEDI-derived global canopy height products in steep Himalayan forests","authors":"Niti B. Mishra , Paras Bikram Singh","doi":"10.1016/j.jag.2026.105203","DOIUrl":"10.1016/j.jag.2026.105203","url":null,"abstract":"<div><div>Accurate canopy-height information in steep mountain forests is essential for biomass estimation, habitat characterization, and monitoring climate-sensitive ecotones, yet the reliability of spaceborne LiDAR height products in rugged terrain remains poorly constrained. Here we benchmarked Global Ecosystem Dynamics Investigation (GEDI) footprint canopy-height metrics and two GEDI-derived global canopy-height products (GFCH 2019 and GCH 2020) using a 1 m UAV-LiDAR canopy height model acquired along a ∼ 2,000 m elevational transect in the central Himalayas, Nepal. We show that GEDI RH95, paired with the UAV P95 reference in a like-with-like percentile comparison, reproduces the overall canopy-height distribution with minimal systematic bias but substantial footprint-scale dispersion, with errors that are spatially structured by terrain and sub-footprint heterogeneity. A footprint-shift experiment indicates that plausible horizontal misregistration contributes comparatively little to overall uncertainty reinforcing that topography and within-footprint variability dominate the error budget in this setting. At the map scale, GFCH 2019 captures broad elevational patterns but systematically underestimates tall canopy heights and exhibits terrain-dependent residual structure, whereas GCH 2020 shows widespread positive bias and spatially patchy coverage over the LiDAR-mapped domain. Together, these results provide a rare UAV-LiDAR benchmark for interpreting GEDI footprint heights and GEDI-derived canopy height models in a data-scarce, high-relief mountain forest, and highlight the need for terrain-aware uncertainty characterization and local calibration when these products are used for biomass, habitat, and treeline/upper-montane applications.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"147 ","pages":"Article 105203"},"PeriodicalIF":8.6,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147278193","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Spatiotemporal dynamics of flood susceptibility under future precipitation variability, population growth, and land cover change 未来降水变率、人口增长和土地覆盖变化下的洪水易感性时空动态

IF 8.6 Q1 REMOTE SENSING

International journal of applied earth observation and geoinformation : ITC journal

Pub Date : 2026-03-01 Epub Date: 2026-02-23 DOI: 10.1016/j.jag.2026.105193

Zahid Ur Rahman , Meimei Zhang , Fang Chen , Safi Ullah , Lei Wang , Zahoor Ahmad , Muhammad Fahad Baqa

Flood risk in mountainous regions is expected to intensify under the compounding effects of climate change, population growth, and land cover changes. However, there is limited understanding of how these interacting factors will shape future flood risk, particularly in the transboundary and ecologically sensitive Kabul River Basin (KRB). The present study addresses this critical gap by assessing the spatiotemporal patterns of projected flood susceptibility in the KRB from 2020 to 2100 under different future scenarios. Future flood susceptibility was predicted using an eXtreme Gradient Boosting (XGBoost) machine learning model with three dynamic and nine static predictors. The findings indicate a significant shift in flood susceptibility over time. Specifically, the areas classified as “Very Highly” susceptible increased from 11.78% in 2020 to 12.17% in 2040, 14.44% in 2060, 13.32% in 2080, and 13.51% by 2100, while the areas classified as “Very Low” susceptibility steadily declined from 66.17% in 2020 to 56.43% by 2100. The XGBoost model showed strong predictive accuracy (AUC: 0.961–0.962) and high cross-temporal consistency across future scenarios (Correlation: 0.75–0.85), confirming its suitability for flood susceptibility assessment. Bootstrap uncertainty analysis further supported its robustness, with mean AUCs of 0.9817–0.9834, very low standard errors (0.0003), and narrow confidence intervals (0.9719–0.9887). These results underscore the need to integrate dynamic environmental and demographic changes into flood management strategies in KRB. The research offers a transferable outline for flood assessment in climate-sensitive mountainous regions. It provides actionable insights for land use planning and climate adaptation policy aimed at reducing future flood impacts.

在气候变化、人口增长和土地覆盖变化的综合影响下，山区洪水风险预计将加剧。然而，对于这些相互作用的因素如何影响未来的洪水风险，特别是在跨界和生态敏感的喀布尔河流域（KRB），人们的理解有限。本研究通过评估2020 - 2100年不同未来情景下KRB预估洪水易感性的时空格局，解决了这一关键缺口。使用极端梯度增强（XGBoost）机器学习模型预测未来洪水敏感性，该模型具有3个动态预测因子和9个静态预测因子。研究结果表明，随着时间的推移，洪水易感性发生了重大变化。其中，“非常高”易感区从2020年的11.78%上升到2040年的12.17%、2060年的14.44%、2080年的13.32%和2100年的13.51%，“非常低”易感区从2020年的66.17%稳步下降到2100年的56.43%。XGBoost模型在未来情景下具有较强的预测精度（AUC: 0.961 ~ 0.962）和高的跨时间一致性（相关系数：0.75 ~ 0.85），证实了其对洪水易感性评价的适用性。Bootstrap不确定性分析进一步支持了其稳健性，平均auc为0.9817-0.9834，标准误差极低（0.0003），置信区间较窄（0.9719-0.9887）。这些结果强调了将动态环境和人口变化纳入KRB洪水管理战略的必要性。该研究为气候敏感山区的洪水评估提供了一个可转移的框架。它为旨在减少未来洪水影响的土地利用规划和气候适应政策提供了可操作的见解。

{"title":"Spatiotemporal dynamics of flood susceptibility under future precipitation variability, population growth, and land cover change","authors":"Zahid Ur Rahman , Meimei Zhang , Fang Chen , Safi Ullah , Lei Wang , Zahoor Ahmad , Muhammad Fahad Baqa","doi":"10.1016/j.jag.2026.105193","DOIUrl":"10.1016/j.jag.2026.105193","url":null,"abstract":"<div><div>Flood risk in mountainous regions is expected to intensify under the compounding effects of climate change, population growth, and land cover changes. However, there is limited understanding of how these interacting factors will shape future flood risk, particularly in the transboundary and ecologically sensitive Kabul River Basin (KRB). The present study addresses this critical gap by assessing the spatiotemporal patterns of projected flood susceptibility in the KRB from 2020 to 2100 under different future scenarios. Future flood susceptibility was predicted using an eXtreme Gradient Boosting (XGBoost) machine learning model with three dynamic and nine static predictors. The findings indicate a significant shift in flood susceptibility over time. Specifically, the areas classified as “Very Highly” susceptible increased from 11.78% in 2020 to 12.17% in 2040, 14.44% in 2060, 13.32% in 2080, and 13.51% by 2100, while the areas classified as “Very Low” susceptibility steadily declined from 66.17% in 2020 to 56.43% by 2100. The XGBoost model showed strong predictive accuracy (AUC: 0.961–0.962) and high cross-temporal consistency across future scenarios (Correlation: 0.75–0.85), confirming its suitability for flood susceptibility assessment. Bootstrap uncertainty analysis further supported its robustness, with mean AUCs of 0.9817–0.9834, very low standard errors (0.0003), and narrow confidence intervals (0.9719–0.9887). These results underscore the need to integrate dynamic environmental and demographic changes into flood management strategies in KRB. The research offers a transferable outline for flood assessment in climate-sensitive mountainous regions. It provides actionable insights for land use planning and climate adaptation policy aimed at reducing future flood impacts.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"147 ","pages":"Article 105193"},"PeriodicalIF":8.6,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147278198","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Time continuous model sequence reconstruction for rice mapping under growth-constrained knowledge priors 生长约束知识先验下水稻图谱的时间连续模型序列重建

IF 8.6 Q1 REMOTE SENSING

International journal of applied earth observation and geoinformation : ITC journal

Pub Date : 2026-03-01 Epub Date: 2026-02-21 DOI: 10.1016/j.jag.2026.105186

Longcai Zhao , Bing Dong , Na Wang , Qiangzi Li

Large-scale rice mapping is crucial for agricultural management and climate change studies but is hampered by challenges such as the spatiotemporal heterogeneity of high-quality data, sample scarcity, and limited model transferability. To overcome these barriers, this study presents a pioneering framework that shifts the paradigm from traditional reconstruction of spectral data to the reconstruction of time-continuous classification model sequences guided by phenological knowledge constraints. Unlike traditional methods that focus on reconstructing raw spectral data, our approach models the continuous evolution of parameters of classification models throughout the growing season. By training discrete Logistic Regression (LR) models regularized by their temporal predecessors, we ensure biophysical realism and link model dynamics directly to rice canopy development. This sparse sequence is then interpolated into a continuous trajectory also under the constraints of rice canopy development. The model-space interpolation enables optimal handling of observations from any arbitrary date and facilitates robust, sample-free model transfer. Long-term validation (1984–2024) and cross-regional rice mapping applications across China, Japan, Italy, and the USA demonstrate that our method achieves high classification accuracy (OA > 0.95) and robust generalization (relative errors < 10%). This study demonstrates that reconstructing model parameter sequences—rather than the raw data itself—is a feasible and powerful strategy for large-scale rice mapping across diverse climatic zones. Beyond rice mapping, the proposed strategy of reconstructing model parameter sequences may offer a transformative perspective for the broader Earth observation community. It demonstrates that when direct high-quality observations are intermittent, modeling the continuous evolution of the classification logic itself under phenological or geographic knowledge is a powerful alternative to data imputation. This conceptual shift provides a scalable solution for monitoring various highly dynamic Earth surface processes, such as forest phenology, and land degradation, where temporal consistency and physical interpretability are paramount.

大规模水稻制图对农业管理和气候变化研究至关重要，但受到高质量数据时空异质性、样本稀缺性和模型可移植性有限等挑战的阻碍。为了克服这些障碍，本研究提出了一个开创性的框架，将传统的光谱数据重建范式转变为在物候知识约束的指导下重建时间连续分类模型序列。与传统方法专注于重建原始光谱数据不同，我们的方法模拟了整个生长季节分类模型参数的持续演变。通过训练离散逻辑回归（LR）模型，我们确保了生物物理的真实性，并将模型动态直接与水稻冠层发育联系起来。然后将该稀疏序列内插到同样受水稻冠层发育约束的连续轨迹中。模型空间插值能够从任何任意日期的观测进行最佳处理，并促进鲁棒，无样本模型转移。长期验证（1984-2024）和中国、日本、意大利和美国的跨区域水稻制图应用表明，我们的方法具有较高的分类精度（OA > 0.95）和稳健的泛化（相对误差<； 10%）。这项研究表明，重建模型参数序列——而不是原始数据本身——是一种可行而有力的策略，可以用于跨不同气候带的大规模水稻制图。除了水稻制图之外，所提出的重建模型参数序列的策略可能为更广泛的地球观测界提供一个变革的视角。它表明，当直接的高质量观测是间歇性的，在物候或地理知识下对分类逻辑本身的连续演变进行建模是数据输入的有力替代方案。这种概念上的转变为监测各种高度动态的地球表面过程提供了一种可扩展的解决方案，例如森林物候和土地退化，在这些过程中，时间一致性和物理可解释性至关重要。

{"title":"Time continuous model sequence reconstruction for rice mapping under growth-constrained knowledge priors","authors":"Longcai Zhao , Bing Dong , Na Wang , Qiangzi Li","doi":"10.1016/j.jag.2026.105186","DOIUrl":"10.1016/j.jag.2026.105186","url":null,"abstract":"<div><div>Large-scale rice mapping is crucial for agricultural management and climate change studies but is hampered by challenges such as the spatiotemporal heterogeneity of high-quality data, sample scarcity, and limited model transferability. To overcome these barriers, this study presents a pioneering framework that shifts the paradigm from traditional reconstruction of spectral data to the reconstruction of time-continuous classification model sequences guided by phenological knowledge constraints. Unlike traditional methods that focus on reconstructing raw spectral data, our approach models the continuous evolution of parameters of classification models throughout the growing season. By training discrete Logistic Regression (LR) models regularized by their temporal predecessors, we ensure biophysical realism and link model dynamics directly to rice canopy development. This sparse sequence is then interpolated into a continuous trajectory also under the constraints of rice canopy development. The model-space interpolation enables optimal handling of observations from any arbitrary date and facilitates robust, sample-free model transfer. Long-term validation (1984–2024) and cross-regional rice mapping applications across China, Japan, Italy, and the USA demonstrate that our method achieves high classification accuracy (OA > 0.95) and robust generalization (relative errors < 10%). This study demonstrates that reconstructing model parameter sequences—rather than the raw data itself—is a feasible and powerful strategy for large-scale rice mapping across diverse climatic zones. Beyond rice mapping, the proposed strategy of reconstructing model parameter sequences may offer a transformative perspective for the broader Earth observation community. It demonstrates that when direct high-quality observations are intermittent, modeling the continuous evolution of the classification logic itself under phenological or geographic knowledge is a powerful alternative to data imputation. This conceptual shift provides a scalable solution for monitoring various highly dynamic Earth surface processes, such as forest phenology, and land degradation, where temporal consistency and physical interpretability are paramount.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"147 ","pages":"Article 105186"},"PeriodicalIF":8.6,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147278200","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Keeping pace with a changing planet: An interactive segmentation framework for refining delineations of dynamic Earth features with the Segment Anything Model 与不断变化的星球保持同步：一个交互式分割框架，用于细化动态地球特征的描述与分段任何模型

IF 8.6 Q1 REMOTE SENSING

International journal of applied earth observation and geoinformation : ITC journal

Pub Date : 2026-03-01 Epub Date: 2026-02-20 DOI: 10.1016/j.jag.2026.105187

Yili Yang , Arav Agarwal , Marlena Bartkus , Heidi Rodenhizer , Wenwen Li , Hyunho Lee , Chia-Yu Hsu , Greg Fiske , Stefano Potter , Anna Liljedahl , Brendan M. Rogers , Susan M. Natali

Accurate delineations of Earth’s surface features are crucial for environmental science, but for dynamic landscapes like thawing permafrost, these datasets rapidly become outdated and inaccurate. Manual refinement of these outdated labels is prohibitively slow, while existing foundation segmentation models often fail to accurately segment complex natural features in satellite images. To address this, we aim to use inaccurate labels as geospatial priors to prompt a vision foundation model for zero-shot semantic segmentation of Earth observation objects with current satellite imagery. We developed an interactive segmentation framework that synergistically combines the Segment Anything Model (SAM) with expert oversight. Our system features two components: (1) a scalable automated pipeline that uses historical delineations as a strong geospatial prior for prompting SAM, and (2) an interactive tool that enables experts to refine outputs and create new, high-fidelity labels rapidly. Using the challenging case study of retrogressive thaw slumps, we demonstrate that this human–AI partnership achieves robust segmentation performance, with a best Intersection over Union of 0.80 ± 0.10 for routine updates. This approach matches the quality of manual expert digitisation but at twice the efficiency. This model- and task-agnostic framework offers a practical, transferable solution for generating the dense, time-series data required to monitor landscape changes, transforming a laborious mapping task into an AI-assisted, expert-overseen workflow.

准确描绘地球表面特征对环境科学至关重要，但对于像永久冻土融化这样的动态景观，这些数据集很快就会过时和不准确。人工对这些过时的标签进行细化的速度非常慢，而现有的基础分割模型往往无法准确分割卫星图像中复杂的自然特征。为了解决这个问题，我们的目标是使用不准确的标签作为地理空间先验，以提示基于当前卫星图像的地球观测对象的零射击语义分割的视觉基础模型。我们开发了一个交互式细分框架，将细分模型（SAM）与专家监督协同结合。我们的系统有两个组成部分：(1)一个可扩展的自动化管道，它使用历史描述作为提示SAM的强大地理空间先验；(2)一个交互式工具，使专家能够精炼输出并快速创建新的高保真度标签。使用具有挑战性的逆行解冻滑坡案例研究，我们证明了这种人类-人工智能合作伙伴关系实现了稳健的分割性能，对于常规更新，最佳交集超过联合为0.80±0.10。这种方法与人工专家数字化的质量相当，但效率是人工专家数字化的两倍。这种与模型和任务无关的框架提供了一种实用的、可转移的解决方案，用于生成监测景观变化所需的密集时间序列数据，将费力的制图任务转化为人工智能辅助的、专家监督的工作流程。

{"title":"Keeping pace with a changing planet: An interactive segmentation framework for refining delineations of dynamic Earth features with the Segment Anything Model","authors":"Yili Yang , Arav Agarwal , Marlena Bartkus , Heidi Rodenhizer , Wenwen Li , Hyunho Lee , Chia-Yu Hsu , Greg Fiske , Stefano Potter , Anna Liljedahl , Brendan M. Rogers , Susan M. Natali","doi":"10.1016/j.jag.2026.105187","DOIUrl":"10.1016/j.jag.2026.105187","url":null,"abstract":"<div><div>Accurate delineations of Earth’s surface features are crucial for environmental science, but for dynamic landscapes like thawing permafrost, these datasets rapidly become outdated and inaccurate. Manual refinement of these outdated labels is prohibitively slow, while existing foundation segmentation models often fail to accurately segment complex natural features in satellite images. To address this, we aim to use inaccurate labels as geospatial priors to prompt a vision foundation model for zero-shot semantic segmentation of Earth observation objects with current satellite imagery. We developed an interactive segmentation framework that synergistically combines the Segment Anything Model (SAM) with expert oversight. Our system features two components: (1) a scalable automated pipeline that uses historical delineations as a strong geospatial prior for prompting SAM, and (2) an interactive tool that enables experts to refine outputs and create new, high-fidelity labels rapidly. Using the challenging case study of retrogressive thaw slumps, we demonstrate that this human–AI partnership achieves robust segmentation performance, with a best Intersection over Union of 0.80 ± 0.10 for routine updates. This approach matches the quality of manual expert digitisation but at twice the efficiency. This model- and task-agnostic framework offers a practical, transferable solution for generating the dense, time-series data required to monitor landscape changes, transforming a laborious mapping task into an AI-assisted, expert-overseen workflow.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"147 ","pages":"Article 105187"},"PeriodicalIF":8.6,"publicationDate":"2026-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147278244","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0