International journal of applied earth observation and geoinformation : ITC journal最新文献_第10页

Assessing socioeconomic and climate driven road maintenance priorities in Southeast Asia using remote sensing approach 利用遥感方法评估东南亚社会经济和气候驱动的道路养护优先事项

IF 8.6 Q1 REMOTE SENSING

International journal of applied earth observation and geoinformation : ITC journal

Pub Date : 2025-12-10 DOI: 10.1016/j.jag.2025.104989

Anjar Dimara Sakti , Muhammad Asa , Agung Budi Harto , Tania Septi Anggraini , Cokro Santoso , Albertus Deliar , Riantini Virtriana , Akhmad Riqqi , Budhy Soeksmantono , Dudy Darmawan Wijaya , Can Trong Nguyen , Khairul Nizam Abdul Maulud , Maya Safira , Ketut Wikantika

Road infrastructure plays a vital role in national and regional development, particularly in Southeast Asia, where rapid economic growth is increasing pressure on transport systems. However, uneven investment, environmental stressors, and limited data-driven tools continue to hinder effective road maintenance planning. Previous studies have utilized remote sensing and statistical models for infrastructure analysis, but the integration of long-term environmental indicators with spatial prioritization methods remains limited. This study addresses this gap by developing a Road Maintenance Priority Index (RMPI) using ten parameters, including nighttime lights, population density, industrial zones, land surface temperature, precipitation, and wind speed. These variables were analyzed through machine learning regression and multi-criteria decision analysis to classify road segments into priority levels. Results show that 45.08 percent of roads fall into the low-priority category, followed by moderate (39.69 percent), high (9.06 percent), and very high (0.88 percent). Countries such as Singapore, Brunei, and Malaysia exhibited the highest RMPI scores, reflecting urgent maintenance needs, while Timor-Leste, Myanmar, and Laos scored lowest. The findings offer a transferable and scalable framework to support evidence-based infrastructure planning in economically and environmentally diverse regions.

道路基础设施在国家和区域发展中发挥着至关重要的作用，特别是在东南亚，那里的快速经济增长正在加大对运输系统的压力。然而，投资不均、环境压力因素和有限的数据驱动工具仍然阻碍着有效的道路养护规划。以往的研究利用遥感和统计模型进行基础设施分析，但将长期环境指标与空间优先排序方法相结合仍然有限。本研究利用夜间灯光、人口密度、工业区、地表温度、降水和风速等10个参数，建立了道路养护优先指数（RMPI），解决了这一问题。通过机器学习回归和多准则决策分析对这些变量进行分析，将路段划分为优先级。结果显示，45.08%的道路属于低优先级，其次是中等（39.69%）、高（9.06%）和非常高（0.88%）。新加坡、文莱和马来西亚等国家的RMPI得分最高，反映了迫切的维护需求，而东帝汶、缅甸和老挝得分最低。研究结果提供了一个可转移和可扩展的框架，以支持经济和环境多样化地区的循证基础设施规划。

{"title":"Assessing socioeconomic and climate driven road maintenance priorities in Southeast Asia using remote sensing approach","authors":"Anjar Dimara Sakti , Muhammad Asa , Agung Budi Harto , Tania Septi Anggraini , Cokro Santoso , Albertus Deliar , Riantini Virtriana , Akhmad Riqqi , Budhy Soeksmantono , Dudy Darmawan Wijaya , Can Trong Nguyen , Khairul Nizam Abdul Maulud , Maya Safira , Ketut Wikantika","doi":"10.1016/j.jag.2025.104989","DOIUrl":"10.1016/j.jag.2025.104989","url":null,"abstract":"<div><div>Road infrastructure plays a vital role in national and regional development, particularly in Southeast Asia, where rapid economic growth is increasing pressure on transport systems. However, uneven investment, environmental stressors, and limited data-driven tools continue to hinder effective road maintenance planning. Previous studies have utilized remote sensing and statistical models for infrastructure analysis, but the integration of long-term environmental indicators with spatial prioritization methods remains limited. This study addresses this gap by developing a Road Maintenance Priority Index (RMPI) using ten parameters, including nighttime lights, population density, industrial zones, land surface temperature, precipitation, and wind speed. These variables were analyzed through machine learning regression and multi-criteria decision analysis to classify road segments into priority levels. Results show that 45.08 percent of roads fall into the low-priority category, followed by moderate (39.69 percent), high (9.06 percent), and very high (0.88 percent). Countries such as Singapore, Brunei, and Malaysia exhibited the highest RMPI scores, reflecting urgent maintenance needs, while Timor-Leste, Myanmar, and Laos scored lowest. The findings offer a transferable and scalable framework to support evidence-based infrastructure planning in economically and environmentally diverse regions.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"146 ","pages":"Article 104989"},"PeriodicalIF":8.6,"publicationDate":"2025-12-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145738279","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

An improved gap probability estimation method accounting for radiometric effects in airborne LiDAR intensity 一种考虑机载激光雷达强度辐射效应的改进间隙概率估计方法

IF 8.6 Q1 REMOTE SENSING

International journal of applied earth observation and geoinformation : ITC journal

Pub Date : 2025-12-10 DOI: 10.1016/j.jag.2025.105010

Lijie Guo, Lei Deng

Gap probability (P) is a key indicator of vegetation canopy structure and can be effectively estimated using intensity data from airborne laser scanning (ALS) point clouds. However, point cloud intensity is highly susceptible to radiometric effects. Even for a specific natural target, its intensity can vary across three-dimensional space, which may reduce the accuracy of P estimation. To address this issue, we propose a novel method for estimating P that corrects the influence of radiometric effects on point cloud intensity (P_{RE_COR}). The method consists of the following main steps: first, laser pulses are classified into vegetation-ground, pure-vegetation, and pure-ground pulses. Then, the intensity of vegetation-ground pulses is corrected using the inverse distance square law and the cosine law of incidence angle. Finally, the corrected intensity values are used to estimate the vegetation-ground reflectance ratio based on a linear relationship between their return energies. This ratio is then used to calculate the canopy P. The proposed method was evaluated using both simulated ALS point cloud data and (National Ecological Observatory Network) NEON ALS point cloud data. The results show that for the simulated data, under varying canopy cover, flight altitudes, and mean scan angles, the proposed method achieved relative root mean square errors (rRMSE) below 5.37%, 5.94%, and 21.51%, and mean absolute errors (MAE) below 0.025, 0.011, and 0.067, respectively. Compared with the traditional P_Fitted method, rRMSE was reduced by up to 1.90%, 1.94%, and 21.20%, and MAE decreased by up to 0.010, 0.003, and 0.096, respectively. For the NEON ALS data, when the scan angle exceeded 20°, the proposed method may improv accuracy by more than 5.39%, with possible MAE improvements exceeding 0.019. Overall, these results demonstrate that correcting radiometric effects on point cloud intensity can substantially enhance both the accuracy and stability of canopy P estimation, with particularly notable benefits under large scan angle conditions.

间隙概率（Gap probability， P）是植被冠层结构的关键指标，利用机载激光扫描（ALS）点云的强度数据可以有效估算植被冠层结构。然而，点云强度极易受到辐射效应的影响。即使对于特定的自然目标，其强度也可能在三维空间中变化，这可能会降低P估计的准确性。为了解决这一问题，我们提出了一种新的估算P值的方法，该方法校正了辐射效应对点云强度（PRE_COR）的影响。该方法主要包括以下几个步骤：首先，将激光脉冲分为植被地面脉冲、纯植被脉冲和纯地面脉冲。然后，利用距离平方反比定律和入射角余弦定律对植被-地面脉冲强度进行校正。最后，利用校正后的强度值，根据植被与地面的反射能量之间的线性关系，估算植被与地面的反射率。利用模拟的ALS点云数据和（National Ecological Observatory Network） NEON ALS点云数据对所提出的方法进行了评估。结果表明，在不同的冠层覆盖度、飞行高度和平均扫描角度下，该方法的相对均方根误差（rRMSE）分别小于5.37%、5.94%和21.51%，平均绝对误差（MAE）分别小于0.025、0.011和0.067。与传统pfitting方法相比，rRMSE分别降低了1.90%、1.94%和21.20%，MAE分别降低了0.010、0.003和0.096。对于NEON ALS数据，当扫描角度大于20°时，所提方法可将精度提高5.39%以上，MAE可能提高0.019以上。总的来说，这些结果表明，校正辐射对点云强度的影响可以大大提高冠层P估计的精度和稳定性，特别是在大扫描角度条件下的效果尤为显著。

{"title":"An improved gap probability estimation method accounting for radiometric effects in airborne LiDAR intensity","authors":"Lijie Guo, Lei Deng","doi":"10.1016/j.jag.2025.105010","DOIUrl":"10.1016/j.jag.2025.105010","url":null,"abstract":"<div><div>Gap probability (P) is a key indicator of vegetation canopy structure and can be effectively estimated using intensity data from airborne laser scanning (ALS) point clouds. However, point cloud intensity is highly susceptible to radiometric effects. Even for a specific natural target, its intensity can vary across three-dimensional space, which may reduce the accuracy of P estimation. To address this issue, we propose a novel method for estimating P that corrects the influence of radiometric effects on point cloud intensity (P<sub>RE_COR</sub>). The method consists of the following main steps: first, laser pulses are classified into vegetation-ground, pure-vegetation, and pure-ground pulses. Then, the intensity of vegetation-ground pulses is corrected using the inverse distance square law and the cosine law of incidence angle. Finally, the corrected intensity values are used to estimate the vegetation-ground reflectance ratio based on a linear relationship between their return energies. This ratio is then used to calculate the canopy P. The proposed method was evaluated using both simulated ALS point cloud data and (National Ecological Observatory Network) NEON ALS point cloud data. The results show that for the simulated data, under varying canopy cover, flight altitudes, and mean scan angles, the proposed method achieved relative root mean square errors (rRMSE) below 5.37%, 5.94%, and 21.51%, and mean absolute errors (MAE) below 0.025, 0.011, and 0.067, respectively. Compared with the traditional P<sub>Fitted</sub> method, rRMSE was reduced by up to 1.90%, 1.94%, and 21.20%, and MAE decreased by up to 0.010, 0.003, and 0.096, respectively. For the NEON ALS data, when the scan angle exceeded 20°, the proposed method may improv accuracy by more than 5.39%, with possible MAE improvements exceeding 0.019. Overall, these results demonstrate that correcting radiometric effects on point cloud intensity can substantially enhance both the accuracy and stability of canopy P estimation, with particularly notable benefits under large scan angle conditions.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"146 ","pages":"Article 105010"},"PeriodicalIF":8.6,"publicationDate":"2025-12-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145738265","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

Clip-based road-marking detection with LLM-guided driving prompts 基于剪辑的道路标记检测与llm引导的驾驶提示

IF 8.6 Q1 REMOTE SENSING

International journal of applied earth observation and geoinformation : ITC journal

Pub Date : 2025-12-08 DOI: 10.1016/j.jag.2025.105012

Shaofan Sheng , Nicolette Formosa , Yuxiang Feng , Mohammed Quddus

The advancement of artificial intelligence (AI) has significantly improved the perception and decision-making abilities of autonomous vehicles (AVs), yet real-time and accurate road marking detection remains difficult under faded markings, nighttime scenes and adverse road–weather conditions. This paper presents RoadGPT, a vision–language pipeline that couples a CLIP-based detector (RoadCLIP) with a planner-facing, LLM-generated advisory layer. The model is trained and evaluated on 3,696 road marking images covering 14 UK Highway Code classes, comprising 2,576 real images (69.7 %) from Google Street View and 1,120 text-to-image synthetic images (30.3 %) that broaden rare appearances and degraded conditions. We use 2,968 images for training (2,072 real, 896 virtual) and 728 images for testing (504 real, 224 virtual), keeping the same 70:30 real–virtual ratio in both splits. On this test set, RoadCLIP attains 98.5 % precision, 97.9 % recall and 98.2 % F1 for non-lane markings, while lane-marking subclasses reach up to 89.8 % F1. The advisory layer transforms recognised markings into structured driving prompts and is assessed via semantic similarity using Sentence-BERT (all-mpnet-base-v2) cosine scores against Highway Code-based references, achieving 89.3 % similarity, alongside an external LLM-as-judge rating of 4.68/5 for accuracy, completeness, and concise effectiveness. The full camera-to-advisory path runs in real time at 135 FPS (batch size 1, 224 × 224) on an RTX 4070 under a unified timing protocol. A remaining limitation is that visually similar lane-marking classes and extreme low-light scenes still reduce discriminability compared with symbol-like, non-lane markings.

人工智能（AI）的进步大大提高了自动驾驶汽车（av）的感知和决策能力，但在褪色的标记、夜间场景和恶劣的道路天气条件下，实时准确的道路标记检测仍然很困难。本文介绍了RoadGPT，一种视觉语言管道，将基于clip的检测器（RoadCLIP）与面向规划器的llm生成的咨询层耦合在一起。该模型在涵盖14个英国公路法规类别的3,696张道路标记图像上进行了训练和评估，其中包括来自谷歌街景的2,576张真实图像（69.7%）和1,120张文本到图像的合成图像（30.3%），这些图像扩展了罕见的外观和退化的条件。我们使用2,968张图像用于训练（2,072张真实图像，896张虚拟图像）和728张图像用于测试（504张真实图像，224张虚拟图像），在两个分割中保持相同的70:30实-虚比例。在这个测试集中，RoadCLIP在非车道标记上达到了98.5%的准确率，97.9%的召回率和98.2%的F1，而车道标记子类达到了89.8%的F1。咨询层将识别的标记转换为结构化的驾驶提示，并使用Sentence-BERT （all-mpnet-base-v2）余弦分数对基于公路法规的参考进行语义相似性评估，达到89.3%的相似性，以及外部llm作为法官的准确性，完整性和简洁有效性评级为4.68/5。完整的摄像机到咨询路径在统一定时协议下的RTX 4070上以135 FPS（批处理大小1,224 × 224）实时运行。剩下的一个限制是，视觉上相似的车道标记类别和极端低光场景仍然会降低与符号式非车道标记相比的可识别性。

{"title":"Clip-based road-marking detection with LLM-guided driving prompts","authors":"Shaofan Sheng , Nicolette Formosa , Yuxiang Feng , Mohammed Quddus","doi":"10.1016/j.jag.2025.105012","DOIUrl":"10.1016/j.jag.2025.105012","url":null,"abstract":"<div><div>The advancement of artificial intelligence (AI) has significantly improved the perception and decision-making abilities of autonomous vehicles (AVs), yet real-time and accurate road marking detection remains difficult under faded markings, nighttime scenes and adverse road–weather conditions. This paper presents RoadGPT, a vision–language pipeline that couples a CLIP-based detector (RoadCLIP) with a planner-facing, LLM-generated advisory layer. The model is trained and evaluated on 3,696 road marking images covering 14 UK Highway Code classes, comprising 2,576 real images (69.7 %) from Google Street View and 1,120 text-to-image synthetic images (30.3 %) that broaden rare appearances and degraded conditions. We use 2,968 images for training (2,072 real, 896 virtual) and 728 images for testing (504 real, 224 virtual), keeping the same 70:30 real–virtual ratio in both splits. On this test set, RoadCLIP attains 98.5 % precision, 97.9 % recall and 98.2 % F1 for non-lane markings, while lane-marking subclasses reach up to 89.8 % F1. The advisory layer transforms recognised markings into structured driving prompts and is assessed via semantic similarity using Sentence-BERT (all-mpnet-base-v2) cosine scores against Highway Code-based references, achieving 89.3 % similarity, alongside an external LLM-as-judge rating of 4.68/5 for accuracy, completeness, and concise effectiveness. The full camera-to-advisory path runs in real time at 135 FPS (batch size 1, 224 × 224) on an RTX 4070 under a unified timing protocol. A remaining limitation is that visually similar lane-marking classes and extreme low-light scenes still reduce discriminability compared with symbol-like, non-lane markings.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"146 ","pages":"Article 105012"},"PeriodicalIF":8.6,"publicationDate":"2025-12-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145738409","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

ALE-CS-NGMS: An advanced approach for individual tree stem- and branch- scale structural parameters extraction using ULS and BLS point clouds ALE-CS-NGMS：一种利用ULS和BLS点云提取单株树干和枝干尺度结构参数的先进方法

IF 8.6 Q1 REMOTE SENSING

International journal of applied earth observation and geoinformation : ITC journal

Pub Date : 2025-12-07 DOI: 10.1016/j.jag.2025.105007

Shanqing Gao , Xin Shen , Caiqin Shen , Lin Cao

Individual tree-level fine-scale (especially stem- and branch- scales) structural parameters constitute a critical foundation for tree structural trait assessments, biomass component estimations and tree physiological property evaluations. However, the Unmanned Aerial Vehicle (UAV) Laser Scanning (ULS) has limitations in sampling distance and penetrating capability in dense tree canopies, thus restricting its ability to extract detailed stem- and branch-scale structural parameters. The emergence of both advanced ULS and Backpack Laser Scanning (BLS) technologies have potential to precisely extract fine-scale structural parameters of individual trees. In this study, we proposed an advanced ALE-CS-NGMS approach for individual tree (Poplar (Populus spp.)) stem- and branch- scale structural parameters extraction by ULS and BLS point clouds. First, an Adaptive Least-squares Ellipse (ALE) fitting algorithm was developed to accurately derive the stem diameter of individual trees. Second, a Canopy-stem Separation (CS) model was built by identifying canopy point cloud through derivatives based on the vertical distribution profile of individual trees, while canopy volume was delineated by the AlphaShape as well as a voxel-based algorithm. Finally, a method integrating Neighborhood Graphs and Minimum Spanning (NGMS) was developed to extract individual tree stem, and stem taper curves were fitted to estimate individual-tree stem volume. The results demonstrated that the developed ALE approach yielded a root mean square error (RMSE) of 2.87 cm, representing an accuracy enhancement approximately 0.47 cm for DBH estimation. The NGMS approach produced RMSEs of 0.33 m³ and 0.40 m³ for stem volume estimation by using BLS and BLS + ULS data. The CS model achieved RMSEs of 6.48 m³ and 3.48 m³ for canopy volume estimation with the BLS and BLS + ULS data, respectively. Branch inclination angles exhibited an increase with stand age, generally ranging between 60° and 100°. The distribution of branch inclination across stands of varying ages revealed that in the 8-year-old and 12-year-old plots, branch angles fell within the 60°-90° interval.

单株精细尺度（特别是茎、枝尺度）结构参数是评价树木结构性状、生物量成分和生理特性的重要基础。然而，无人机（UAV）激光扫描（ULS）在密集树冠的采样距离和穿透能力方面存在局限性，从而限制了其提取枝干尺度结构参数的能力。先进的ULS和双肩包激光扫描（BLS）技术的出现，都有可能精确提取单个树木的精细结构参数。在这项研究中，我们提出了一种先进的ALE-CS-NGMS方法，用于利用ULS和BLS点云提取单树（杨树（Populus spp.））的茎和枝尺度结构参数。首先，提出了一种自适应最小二乘椭圆（ALE）拟合算法，精确地推导出单株树的茎粗；其次，基于单株树木垂直分布轮廓，通过导数识别冠层点云，建立冠层-茎分离（canopy -stem Separation， CS）模型，同时利用AlphaShape和基于体素的算法圈定冠层体积；最后，提出了结合邻域图和最小生成（NGMS）的方法提取单株树干，并拟合树干锥度曲线估算单株树干体积。结果表明，该方法的均方根误差（RMSE）为2.87 cm，提高了约0.47 cm的胸径估计精度。NGMS方法使用BLS和BLS + ULS数据估算茎干体积的均方根误差分别为0.33 m3和0.40 m3。CS模型对BLS和BLS + ULS数据的冠层体积估算均方根误差分别为6.48 m3和3.48 m3。枝条倾角随林龄增加而增大，一般在60°~ 100°之间。不同林龄林分枝条倾角分布表明，8年和12年林分枝条倾角分布在60°~ 90°区间。

{"title":"ALE-CS-NGMS: An advanced approach for individual tree stem- and branch- scale structural parameters extraction using ULS and BLS point clouds","authors":"Shanqing Gao , Xin Shen , Caiqin Shen , Lin Cao","doi":"10.1016/j.jag.2025.105007","DOIUrl":"10.1016/j.jag.2025.105007","url":null,"abstract":"<div><div>Individual tree-level fine-scale (especially stem- and branch- scales) structural parameters constitute a critical foundation for tree structural trait assessments, biomass component estimations and tree physiological property evaluations. However, the Unmanned Aerial Vehicle (UAV) Laser Scanning (ULS) has limitations in sampling distance and penetrating capability in dense tree canopies, thus restricting its ability to extract detailed stem- and branch-scale structural parameters. The emergence of both advanced ULS and Backpack Laser Scanning (BLS) technologies have potential to precisely extract fine-scale structural parameters of individual trees. In this study, we proposed an advanced ALE-CS-NGMS approach for individual tree (<em>Poplar</em> (<em>Populus</em> spp.)) stem- and branch- scale structural parameters extraction by ULS and BLS point clouds. First, an Adaptive Least-squares Ellipse (ALE) fitting algorithm was developed to accurately derive the stem diameter of individual trees. Second, a Canopy-stem Separation (CS) model was built by identifying canopy point cloud through derivatives based on the vertical distribution profile of individual trees, while canopy volume was delineated by the AlphaShape as well as a voxel-based algorithm. Finally, a method integrating Neighborhood Graphs and Minimum Spanning (NGMS) was developed to extract individual tree stem, and stem taper curves were fitted to estimate individual-tree stem volume. The results demonstrated that the developed ALE approach yielded a root mean square error (RMSE) of 2.87 cm, representing an accuracy enhancement approximately 0.47 cm for DBH estimation. The NGMS approach produced RMSEs of 0.33 m<sup>3</sup> and 0.40 m<sup>3</sup> for stem volume estimation by using BLS and BLS + ULS data. The CS model achieved RMSEs of 6.48 m<sup>3</sup> and 3.48 m<sup>3</sup> for canopy volume estimation with the BLS and BLS + ULS data, respectively. Branch inclination angles exhibited an increase with stand age, generally ranging between 60° and 100°. The distribution of branch inclination across stands of varying ages revealed that in the 8-year-old and 12-year-old plots, branch angles fell within the 60°-90° interval.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"146 ","pages":"Article 105007"},"PeriodicalIF":8.6,"publicationDate":"2025-12-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145697394","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

Hybrid wetland city map: Improved wetland characterization through the synergy of global land cover products 混合湿地城市地图：通过全球土地覆盖产品的协同作用改善湿地特征

IF 8.6 Q1 REMOTE SENSING

International journal of applied earth observation and geoinformation : ITC journal

Pub Date : 2025-12-06 DOI: 10.1016/j.jag.2025.104994

Xiaogan Yin , Weiguo Jiang , Zhe Yang , Ziyan Ling , Nandin-Erdene Tsendbazar , Peng Hou , Yue Deng , Xiaoya Wang , Zhijie Xiao , Xiao Li , Miaolong Lin

Global wetlands are experiencing severe degradation due to climate change and human activities. Under the Ramsar Convention, the Wetland City Accreditation promotes cities to protect and sustainably manage their urban wetlands. The accreditation system was launched in 2015. To date, 43 cities worldwide have obtained this certification, whose dynamic assessment depends on precise mapping of land use and wetlands. Existing global land cover datasets often show low accuracy in identifying wetlands and limited capacity to characterize wetland types within urban areas. we developed a hybrid Wetland City Map (WCM), by fusing three global 10 m-resolution products: Dynamic World, ESA WorldCover, and ESRI Land Cover. We applied a Weighted Voting and Knowledge-based Decision Rule method to achieve this fusion. This method overcomes the limitations of the input datasets by combining their complementary strengths to improve overall wetland classification and by applying expert-derived rules to enhance the delineation of wetland types within cities. The WCM achieves an average overall accuracy of 86.93 % and a kappa of 0.825. In all cities, its accuracy surpasses the three land cover products by 2 %-26 %. The visual comparison shows WCM performs better in wetland classification and spatial detail, with F1 scores of 90.33 % (water), 64.09 % (marsh), 71.67 % (tidal flat/flooded flat), and 92.17 % (mangrove). It more accurately reflects wetland coverage and changes. Wetland coverage varies across cities, with higher coverage in Asia and lower in Europe and Africa. Individual cities experienced a maximum increase of 6.5 % and decrease of 1.3 % from 2020 to 2021.The WCM supports wetland monitoring, city accreditation, and research aligned with the Ramsar Strategic Plan and Sustainable Development Goals.

由于气候变化和人类活动，全球湿地正在经历严重退化。根据《拉姆萨尔公约》，湿地城市认证旨在促进城市保护和可持续管理其城市湿地。认证制度于2015年启动。迄今为止，全世界有43个城市获得了这一认证，其动态评估依赖于精确的土地利用和湿地地图。现有的全球土地覆盖数据在识别湿地方面往往表现出较低的准确性，并且表征城市地区湿地类型的能力有限。通过融合三个全球10米分辨率的产品：Dynamic World、ESA WorldCover和ESRI Land Cover，我们开发了一个混合型湿地城市地图（WCM）。我们采用加权投票和基于知识的决策规则方法来实现这种融合。该方法通过结合输入数据集的互补优势来改进整体湿地分类，并通过应用专家导出的规则来增强城市内湿地类型的划分，从而克服了输入数据集的局限性。WCM的平均整体准确率为86.93%，kappa为0.825。在所有城市中，其精度比三种土地覆盖产品高出2% - 26%。视觉对比表明，WCM在湿地分类和空间细节方面表现较好，F1得分分别为90.33%（水）、64.09%（沼泽）、71.67%（潮滩/淹滩）和92.17%（红树林）。它更准确地反映了湿地的覆盖和变化。不同城市的湿地覆盖率各不相同，亚洲的覆盖率较高，欧洲和非洲的覆盖率较低。从2020年到2021年，单个城市的最高增长率为6.5%，下降1.3%。WCM支持湿地监测、城市认证和与拉姆萨尔战略规划和可持续发展目标相一致的研究。

{"title":"Hybrid wetland city map: Improved wetland characterization through the synergy of global land cover products","authors":"Xiaogan Yin , Weiguo Jiang , Zhe Yang , Ziyan Ling , Nandin-Erdene Tsendbazar , Peng Hou , Yue Deng , Xiaoya Wang , Zhijie Xiao , Xiao Li , Miaolong Lin","doi":"10.1016/j.jag.2025.104994","DOIUrl":"10.1016/j.jag.2025.104994","url":null,"abstract":"<div><div>Global wetlands are experiencing severe degradation due to climate change and human activities. Under the Ramsar Convention, the Wetland City Accreditation promotes cities to protect and sustainably manage their urban wetlands. The accreditation system was launched in 2015. To date, 43 cities worldwide have obtained this certification, whose dynamic assessment depends on precise mapping of land use and wetlands. Existing global land cover datasets often show low accuracy in identifying wetlands and limited capacity to characterize wetland types within urban areas. we developed a hybrid Wetland City Map (WCM), by fusing three global 10 m-resolution products: Dynamic World, ESA WorldCover, and ESRI Land Cover. We applied a Weighted Voting and Knowledge-based Decision Rule method to achieve this fusion. This method overcomes the limitations of the input datasets by combining their complementary strengths to improve overall wetland classification and by applying expert-derived rules to enhance the delineation of wetland types within cities. The WCM achieves an average overall accuracy of 86.93 % and a kappa of 0.825. In all cities, its accuracy surpasses the three land cover products by 2 %-26 %. The visual comparison shows WCM performs better in wetland classification and spatial detail, with F1 scores of 90.33 % (water), 64.09 % (marsh), 71.67 % (tidal flat/flooded flat), and 92.17 % (mangrove). It more accurately reflects wetland coverage and changes. Wetland coverage varies across cities, with higher coverage in Asia and lower in Europe and Africa. Individual cities experienced a maximum increase of 6.5 % and decrease of 1.3 % from 2020 to 2021.The WCM supports wetland monitoring, city accreditation, and research aligned with the Ramsar Strategic Plan and Sustainable Development Goals.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"146 ","pages":"Article 104994"},"PeriodicalIF":8.6,"publicationDate":"2025-12-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145685752","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

Generative models for SAR–optical image translation: A systematic review sar光学图像翻译的生成模型：系统综述

IF 8.6 Q1 REMOTE SENSING

International journal of applied earth observation and geoinformation : ITC journal

Pub Date : 2025-12-04 DOI: 10.1016/j.jag.2025.105009

Zhao Wang , Zheng Zhang , Xiaojun Shan , Hong-an Wei , Ping Tang

Growing demands in sustainable development and resource management are driving increasing reliance on remote sensing-based Earth observation and image interpretation. In parallel, multimodal collaborative processing is attracting research attention. Synthetic aperture radar (SAR) and optical images offer complementary advantages but pose challenges for simultaneous use due to platform constraints and environmental conditions, often leaving only one modality available and impeding joint analysis. Generative models, particularly generative adversarial networks (GANs) and diffusion models (DMs), address this by learning cross-modal mappings. Translated images preserve structure and semantics while adopting target characteristics, thereby facilitating collaborative use. This review systematically categorizes translation frameworks spanning GANs, DMs, and other generative models. It then details downstream tasks supported by SAR–optical translation, including cloud removal, change detection, semantic segmentation, registration, and object detection, highlighting how translation bridges data gaps and enhances interpretation robustness. Furthermore, we provide open-source code and public datasets, discuss current challenges, and outline future research directions.

在可持续发展和资源管理方面日益增长的需求正在推动越来越多地依赖基于遥感的地球观测和图像判读。与此同时，多模态协同处理也引起了人们的关注。合成孔径雷达（SAR）和光学图像具有互补的优势，但由于平台的限制和环境条件的限制，同时使用也面临挑战，通常只有一种模式可用，阻碍了联合分析。生成模型，特别是生成对抗网络（gan）和扩散模型（dm），通过学习跨模态映射来解决这个问题。翻译后的图像在采用目标特征的同时保留了结构和语义，从而促进了协作使用。本文对跨gan、dm和其他生成模型的翻译框架进行了系统分类。然后详细介绍了sar光学翻译支持的下游任务，包括云移除、变化检测、语义分割、注册和对象检测，重点介绍了翻译如何弥合数据差距并增强解释的鲁棒性。此外，我们提供开源代码和公共数据集，讨论当前的挑战，并概述未来的研究方向。

{"title":"Generative models for SAR–optical image translation: A systematic review","authors":"Zhao Wang , Zheng Zhang , Xiaojun Shan , Hong-an Wei , Ping Tang","doi":"10.1016/j.jag.2025.105009","DOIUrl":"10.1016/j.jag.2025.105009","url":null,"abstract":"<div><div>Growing demands in sustainable development and resource management are driving increasing reliance on remote sensing-based Earth observation and image interpretation. In parallel, multimodal collaborative processing is attracting research attention. Synthetic aperture radar (SAR) and optical images offer complementary advantages but pose challenges for simultaneous use due to platform constraints and environmental conditions, often leaving only one modality available and impeding joint analysis. Generative models, particularly generative adversarial networks (GANs) and diffusion models (DMs), address this by learning cross-modal mappings. Translated images preserve structure and semantics while adopting target characteristics, thereby facilitating collaborative use. This review systematically categorizes translation frameworks spanning GANs, DMs, and other generative models. It then details downstream tasks supported by SAR–optical translation, including cloud removal, change detection, semantic segmentation, registration, and object detection, highlighting how translation bridges data gaps and enhances interpretation robustness. Furthermore, we provide open-source code and public datasets, discuss current challenges, and outline future research directions.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"146 ","pages":"Article 105009"},"PeriodicalIF":8.6,"publicationDate":"2025-12-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145685754","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

Assessment of three remote sensing methods for estimating actual evapotranspiration in a Mediterranean region 估算地中海地区实际蒸散量的三种遥感方法的评估

IF 8.6 Q1 REMOTE SENSING

International journal of applied earth observation and geoinformation : ITC journal

Pub Date : 2025-12-04 DOI: 10.1016/j.jag.2025.105003

Luca Fibbi , Marta Chiesi , Maurizio Pieri , Giorgio Bartolini , Daniele Grifoni , Bernardo Gozzini , Fabio Maselli

A semi-empirical method based on ancillary and remotely sensed data, NDVI-Cws, is currently applied over the Tuscany region to yield daily actual evapotranspiration (ETa) estimates at moderate spatial resolution (250 m) for a 20-year period (2005–2024). The outcome of this exercise is then statistically analysed in comparison with two similar ETa products provided by the MODIS and LSA SAF systems at 500 m and 5 km resolutions, respectively. The analysis relies on the triple collocation strategy and comprises the following two steps: i) examination and inter-comparison of the spatial and temporal ETa variations which occur in 12 areas of the region representative of different climatic conditions and biome types; ii) repetition of these operations at pixel level. The experimental results indicate the existence of clear spatial and temporal ETa variations over most of the region which are differently represented by the three ETa products. The MODIS ETa estimates are significantly higher and lower than those of the other two products for forest and non-forest areas, respectively; the ETa trends estimated by MODIS are poorly concordant with those of the other products, particularly for forests. Consequently, the MODIS ETa estimates reflect only marginally the ETa increases which are evidenced by the other two methods over most of the region. Out of these methods, NDVI-Cws allows a more spatially detailed prediction of the local ETa variability depending on the NDVI dataset used. The implications and consequences of these findings are finally discussed, together with the main future research prospects.

基于辅助和遥感数据的半经验方法NDVI-Cws目前在托斯卡纳地区应用，以中等空间分辨率（250 m）估算20年（2005-2024年）的每日实际蒸散（ETa）。然后，将这一工作的结果与MODIS和LSA SAF系统分别在500米和5公里分辨率下提供的两种类似的ETa产品进行统计分析。该分析依赖于三重搭配策略，包括以下两个步骤：1)对该地区12个具有不同气候条件和生物群系类型的地区的时空ETa变化进行检查和相互比较；Ii)在像素级重复这些操作。实验结果表明，在大部分地区存在明显的ETa时空变化，三种ETa产品所代表的时空变化不同。在森林和非森林地区，MODIS的ETa估计值分别显著高于和低于其他两种产品；MODIS估计的ETa趋势与其他产品，特别是森林产品的趋势不太一致。因此，MODIS的ETa估计值只略微反映了ETa的增加，而其他两种方法在大部分地区都证明了这一点。在这些方法中，NDVI- cws可以根据使用的NDVI数据集对当地ETa变化进行更详细的空间预测。最后讨论了这些发现的意义和后果，并展望了未来的主要研究前景。

{"title":"Assessment of three remote sensing methods for estimating actual evapotranspiration in a Mediterranean region","authors":"Luca Fibbi , Marta Chiesi , Maurizio Pieri , Giorgio Bartolini , Daniele Grifoni , Bernardo Gozzini , Fabio Maselli","doi":"10.1016/j.jag.2025.105003","DOIUrl":"10.1016/j.jag.2025.105003","url":null,"abstract":"<div><div>A semi-empirical method based on ancillary and remotely sensed data, NDVI-Cws, is currently applied over the Tuscany region to yield daily actual evapotranspiration (ETa) estimates at moderate spatial resolution (250 m) for a 20-year period (2005–2024). The outcome of this exercise is then statistically analysed in comparison with two similar ETa products provided by the MODIS and LSA SAF systems at 500 m and 5 km resolutions, respectively. The analysis relies on the triple collocation strategy and comprises the following two steps: i) examination and inter-comparison of the spatial and temporal ETa variations which occur in 12 areas of the region representative of different climatic conditions and biome types; ii) repetition of these operations at pixel level. The experimental results indicate the existence of clear spatial and temporal ETa variations over most of the region which are differently represented by the three ETa products. The MODIS ETa estimates are significantly higher and lower than those of the other two products for forest and non-forest areas, respectively; the ETa trends estimated by MODIS are poorly concordant with those of the other products, particularly for forests. Consequently, the MODIS ETa estimates reflect only marginally the ETa increases which are evidenced by the other two methods over most of the region. Out of these methods, NDVI-Cws allows a more spatially detailed prediction of the local ETa variability depending on the NDVI dataset used. The implications and consequences of these findings are finally discussed, together with the main future research prospects.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"146 ","pages":"Article 105003"},"PeriodicalIF":8.6,"publicationDate":"2025-12-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145659160","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

InSAR and GAT-LSTM integration for dam displacement prediction: Lessons from the Oldman River Dam, Canada InSAR和GAT-LSTM集成用于大坝位移预测：来自加拿大奥德曼河大坝的经验

IF 8.6 Q1 REMOTE SENSING

International journal of applied earth observation and geoinformation : ITC journal

Pub Date : 2025-12-04 DOI: 10.1016/j.jag.2025.104968

Ramin Farhadiani , Sayyed Mohammad Javad Mirzadeh , Ehsan Roshani , Daniel Cusson , Saeid Homayouni

The precise prediction of dam deformation is essential for ensuring infrastructure safety and mitigating geohazards, particularly in regions characterized by limited monitoring studies. This research concentrates on the Oldman River Dam in Alberta, Canada, where Interferometric Synthetic Aperture Radar (InSAR)-based deformation monitoring and prediction remain inadequately explored. A novel framework that integrates a Graph Attention Network with Long Short-Term Memory (GAT-LSTM) has been developed to address the limitations of existing methods, which neglect spatial dependencies among InSAR-derived points and the increased model complexity stemming from point clustering or InSAR time series decomposition. Sentinel-1 data from three passes were processed utilizing a full-resolution InSAR technique, resulting in semi-vertical deformation velocities that demonstrated consistent subsidence along the dam crest, with rates fluctuating from 5.08 to 6.23 mm/yr. A robust correlation between deformation and reservoir water levels was noted, with accelerated crest deformation during the 2017–2019 drawdown period and a potential risk identified due to a significant decline in water levels projected for 2023–2024. The GAT-LSTM model, which captures both spatial and temporal dynamics, outperformed the standard LSTM, achieving 83.64% accurate points compared to 76.90% for the LSTM in short-term forecasting, exhibiting notable reliability along the crest. The peak performance was observed on September 9, 2021, with a Root Mean Square Error of 0.30 ± 0.013 mm and a Mean Absolute Error of 0.22 ± 0.012 mm. The proposed framework would enhance dam safety monitoring by providing actionable short-term predictions, demonstrating potential transferability to other slow-moving infrastructure.

大坝变形的精确预测对于确保基础设施安全和减轻地质灾害至关重要，特别是在监测研究有限的地区。本研究以加拿大阿尔伯塔省的奥德曼河大坝为研究对象，该大坝基于干涉合成孔径雷达（InSAR）的变形监测和预测尚未得到充分探索。为了解决现有方法忽略InSAR衍生点之间的空间依赖性以及由点聚类或InSAR时间序列分解引起的模型复杂性增加的局限性，开发了一种集成了长短期记忆的图注意网络（GAT-LSTM）框架。利用全分辨率InSAR技术处理了来自三次通道的Sentinel-1数据，得到了半垂直变形速度，表明沿坝顶持续下沉，速率在5.08至6.23 mm/年之间波动。变形与水库水位之间存在很强的相关性，在2017-2019年的下降期间，峰值变形加速，并且由于预计2023-2024年的水位显著下降，确定了潜在的风险。GAT-LSTM模型兼顾了时空动态，短期预报准确率为83.64%，优于标准LSTM模型的76.90%，且沿波峰方向具有显著的可靠性。在2021年9月9日达到峰值，均方根误差为0.30±0.013 mm，平均绝对误差为0.22±0.012 mm。拟议的框架将通过提供可操作的短期预测来加强大坝安全监测，并展示潜在的可转移性，用于其他运行缓慢的基础设施。

{"title":"InSAR and GAT-LSTM integration for dam displacement prediction: Lessons from the Oldman River Dam, Canada","authors":"Ramin Farhadiani , Sayyed Mohammad Javad Mirzadeh , Ehsan Roshani , Daniel Cusson , Saeid Homayouni","doi":"10.1016/j.jag.2025.104968","DOIUrl":"10.1016/j.jag.2025.104968","url":null,"abstract":"<div><div>The precise prediction of dam deformation is essential for ensuring infrastructure safety and mitigating geohazards, particularly in regions characterized by limited monitoring studies. This research concentrates on the Oldman River Dam in Alberta, Canada, where Interferometric Synthetic Aperture Radar (InSAR)-based deformation monitoring and prediction remain inadequately explored. A novel framework that integrates a Graph Attention Network with Long Short-Term Memory (GAT-LSTM) has been developed to address the limitations of existing methods, which neglect spatial dependencies among InSAR-derived points and the increased model complexity stemming from point clustering or InSAR time series decomposition. Sentinel-1 data from three passes were processed utilizing a full-resolution InSAR technique, resulting in semi-vertical deformation velocities that demonstrated consistent subsidence along the dam crest, with rates fluctuating from 5.08 to 6.23 mm/yr. A robust correlation between deformation and reservoir water levels was noted, with accelerated crest deformation during the 2017–2019 drawdown period and a potential risk identified due to a significant decline in water levels projected for 2023–2024. The GAT-LSTM model, which captures both spatial and temporal dynamics, outperformed the standard LSTM, achieving 83.64% accurate points compared to 76.90% for the LSTM in short-term forecasting, exhibiting notable reliability along the crest. The peak performance was observed on September 9, 2021, with a Root Mean Square Error of 0.30 ± 0.013 mm and a Mean Absolute Error of 0.22 ± 0.012 mm. The proposed framework would enhance dam safety monitoring by providing actionable short-term predictions, demonstrating potential transferability to other slow-moving infrastructure.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"146 ","pages":"Article 104968"},"PeriodicalIF":8.6,"publicationDate":"2025-12-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145685753","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

A global analysis of SAR altimetry signals over different landcover types 不同地表覆盖类型SAR测高信号的全球分析

IF 8.6 Q1 REMOTE SENSING

International journal of applied earth observation and geoinformation : ITC journal

Pub Date : 2025-12-03 DOI: 10.1016/j.jag.2025.105000

Maximilian Eitel, Michael Schmitt

Satellite radar altimeters were originally designed for water applications, but their echoes over land surfaces remain less well understood. In this study we analyze how Sentinel-3 (S3) synthetic aperture radar (SAR) altimetry waveforms respond to different surface types and what physical characteristics are encoded in the signal. To probe this, we conduct classification experiments with a feature-enhanced one-dimensional convolutional neural network (1D-CNN) and analyze its performance. Since surface type information is relevant for climate, hydrology, and biodiversity applications, understanding these signal responses shows to what extent altimetric waveforms may provide consistent class-specific information despite their large elliptical footprint and heterogeneous landscapes. This study investigates the response of Sentinel-3 altimetry waveforms to different land cover types by employing a 1D-CNN to extract land cover information, complemented by a visual analysis of waveform patterns in relation to surface structures. Our results show that information about the underlying surface is embedded in the signals and can be extracted. They further reveal the sensitivity of Sentinel-3 altimetry to variations in land cover. By enhancing our 1D-CNN model with shape-based and contextual features, it effectively captures surface characteristics despite the large altimeter footprint. An ablation study highlights the complementary role of these features, as their removal negatively impacts performance. The best-performing 1D-CNN achieves a macro-averaged F1 (Macro-F1) score of 0.57 and an overall accuracy of 0.67, outperforming both a random forest and a dummy baseline. The classification includes six surface types: Tree, Shrub, Grass, Crop, Bare/Sparse Vegetation, and Water. Although some misclassification occurs, particularly in transition zones and among classes with similar vegetation structures and soil properties, the model provides valuable insights into systematic waveform behavior, highlighting the potential of SAR altimetry signals to capture broad surface characteristics.

卫星雷达高度计最初是为水上应用而设计的，但它们在陆地表面的回波仍然不太清楚。在本研究中，我们分析了Sentinel-3 （S3）合成孔径雷达（SAR）测高波形对不同地表类型的响应，以及信号中编码的物理特征。为了探讨这一点，我们使用特征增强的一维卷积神经网络（1D-CNN）进行分类实验，并分析其性能。由于地表类型信息与气候、水文和生物多样性应用相关，了解这些信号响应表明，尽管高海拔波形具有较大的椭圆足迹和异质性景观，但它们在多大程度上可以提供一致的类别特定信息。本研究通过使用1D-CNN提取土地覆盖信息，并辅以与地表结构相关的波形模式的可视化分析，研究了Sentinel-3高度计波形对不同土地覆盖类型的响应。我们的研究结果表明，下垫面的信息被嵌入到信号中，并且可以被提取出来。它们进一步揭示了哨兵3号测高对土地覆盖变化的敏感性。通过使用基于形状和上下文的特征增强我们的1D-CNN模型，尽管高度计占用空间很大，但它仍能有效地捕获表面特征。消融研究强调了这些特征的互补作用，因为它们的去除会对性能产生负面影响。表现最好的1D-CNN的宏观平均F1 （Macro-F1）得分为0.57，总体精度为0.67，优于随机森林和虚拟基线。分类包括六种表面类型：树木、灌木、草、作物、裸/稀疏植被和水。虽然会出现一些错误分类，特别是在过渡带和具有相似植被结构和土壤性质的类别之间，但该模型提供了对系统波形行为的有价值的见解，突出了SAR测高信号捕捉广泛地表特征的潜力。

{"title":"A global analysis of SAR altimetry signals over different landcover types","authors":"Maximilian Eitel, Michael Schmitt","doi":"10.1016/j.jag.2025.105000","DOIUrl":"10.1016/j.jag.2025.105000","url":null,"abstract":"<div><div>Satellite radar altimeters were originally designed for water applications, but their echoes over land surfaces remain less well understood. In this study we analyze how Sentinel-3 (S3) synthetic aperture radar (SAR) altimetry waveforms respond to different surface types and what physical characteristics are encoded in the signal. To probe this, we conduct classification experiments with a feature-enhanced one-dimensional convolutional neural network (1D-CNN) and analyze its performance. Since surface type information is relevant for climate, hydrology, and biodiversity applications, understanding these signal responses shows to what extent altimetric waveforms may provide consistent class-specific information despite their large elliptical footprint and heterogeneous landscapes. This study investigates the response of Sentinel-3 altimetry waveforms to different land cover types by employing a 1D-CNN to extract land cover information, complemented by a visual analysis of waveform patterns in relation to surface structures. Our results show that information about the underlying surface is embedded in the signals and can be extracted. They further reveal the sensitivity of Sentinel-3 altimetry to variations in land cover. By enhancing our 1D-CNN model with shape-based and contextual features, it effectively captures surface characteristics despite the large altimeter footprint. An ablation study highlights the complementary role of these features, as their removal negatively impacts performance. The best-performing 1D-CNN achieves a macro-averaged F1 (Macro-F1) score of 0.57 and an overall accuracy of 0.67, outperforming both a random forest and a dummy baseline. The classification includes six surface types: Tree, Shrub, Grass, Crop, Bare/Sparse Vegetation, and Water. Although some misclassification occurs, particularly in transition zones and among classes with similar vegetation structures and soil properties, the model provides valuable insights into systematic waveform behavior, highlighting the potential of SAR altimetry signals to capture broad surface characteristics.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"146 ","pages":"Article 105000"},"PeriodicalIF":8.6,"publicationDate":"2025-12-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145659109","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

Unsupervised multitemporal SAR image change detection via foreground-background collaborative optimization 基于前景-背景协同优化的无监督多时相SAR图像变化检测

IF 8.6 Q1 REMOTE SENSING

International journal of applied earth observation and geoinformation : ITC journal

Pub Date : 2025-12-03 DOI: 10.1016/j.jag.2025.105008

Weisong Li , Yinwei Li , Yiming Zhu , Haipeng Wang

Multitemporal Synthetic Aperture Radar (SAR) image change detection (CD) represents a significant focus in remote sensing interpretation research. Recently, matrix low-rank decomposition theory has gained popularity in this field to exploit inherent structural information without requiring annotated data. However, existing approaches predominantly rely on an idealized assumption that defines changed regions as spatially localized and sparse. This assumption introduces critical theoretical limitations: its sensitivity to change scales results in sparsity constraints failing to characterize large-scale continuous changes and accumulating decomposition errors, while the neglect of low-rank coupling between changed regions and backgrounds further undermines theoretical completeness. To address these issues, we propose a Foreground and Background dual-path collaborative optimization CD framework, namely FBCD. Specifically, a foreground change saliency model is constructed under generalized low-rank constraints, integrating low-rank consistency and local correlation mechanisms to capture complex change patterns. In addition, a background stability model based on low-rank self-representation learning achieves precise background separation through multi-view consistency constraint. Once generating reconstructed difference map, a self-supervised graph-optimized label propagation algorithm is designed to transform binary classification into a graph partitioning optimization problem, which further improves the CD accuracy. Extensive experiments on seven bitemporal benchmark datasets validate the superiority of the proposed method: Compared to state-of-the-art approaches, it achieves average Kappa coefficient improvements of 2.50% for large-scale continuous changes and 4.48% for small-scale localized complex changes. Furthermore, the method also shows strong applicability in short-term time series datasets. The source code will be made available at https://github.com/95xiaoli/FBCD.

多时相合成孔径雷达（SAR）图像变化检测是遥感解译研究的一个重要热点。近年来，矩阵低秩分解理论在该领域得到了广泛的应用，它可以在不需要标注数据的情况下挖掘固有的结构信息。然而，现有的方法主要依赖于一个理想化的假设，即将变化的区域定义为空间局部化和稀疏的。这一假设引入了关键的理论局限性：它对变化尺度的敏感性导致稀疏性约束无法表征大尺度连续变化和累积分解误差，而忽略变化区域和背景之间的低秩耦合进一步破坏了理论的完整性。为了解决这些问题，我们提出了一个前景和后台双路径协同优化CD框架，即FBCD。具体而言，在广义低秩约束下构建前景变化显著性模型，整合低秩一致性和局部相关机制来捕捉复杂的变化模式。此外，基于低秩自表示学习的背景稳定性模型通过多视图一致性约束实现了精确的背景分离。在生成重构差分图后，设计了一种自监督图优化标签传播算法，将二值分类转化为图划分优化问题，进一步提高了CD的准确率。在7个双时间基准数据集上的大量实验验证了该方法的优越性：与现有方法相比，对于大规模连续变化，该方法的Kappa系数平均提高了2.50%，对于小规模局部复杂变化，该方法的Kappa系数平均提高了4.48%。此外，该方法在短期时间序列数据集上也显示出较强的适用性。源代码将在https://github.com/95xiaoli/FBCD上提供。

{"title":"Unsupervised multitemporal SAR image change detection via foreground-background collaborative optimization","authors":"Weisong Li , Yinwei Li , Yiming Zhu , Haipeng Wang","doi":"10.1016/j.jag.2025.105008","DOIUrl":"10.1016/j.jag.2025.105008","url":null,"abstract":"<div><div>Multitemporal Synthetic Aperture Radar (SAR) image change detection (CD) represents a significant focus in remote sensing interpretation research. Recently, matrix low-rank decomposition theory has gained popularity in this field to exploit inherent structural information without requiring annotated data. However, existing approaches predominantly rely on an idealized assumption that defines changed regions as spatially localized and sparse. This assumption introduces critical theoretical limitations: its sensitivity to change scales results in sparsity constraints failing to characterize large-scale continuous changes and accumulating decomposition errors, while the neglect of low-rank coupling between changed regions and backgrounds further undermines theoretical completeness. To address these issues, we propose a Foreground and Background dual-path collaborative optimization CD framework, namely FBCD. Specifically, a foreground change saliency model is constructed under generalized low-rank constraints, integrating low-rank consistency and local correlation mechanisms to capture complex change patterns. In addition, a background stability model based on low-rank self-representation learning achieves precise background separation through multi-view consistency constraint. Once generating reconstructed difference map, a self-supervised graph-optimized label propagation algorithm is designed to transform binary classification into a graph partitioning optimization problem, which further improves the CD accuracy. Extensive experiments on seven bitemporal benchmark datasets validate the superiority of the proposed method: Compared to state-of-the-art approaches, it achieves average Kappa coefficient improvements of 2.50% for large-scale continuous changes and 4.48% for small-scale localized complex changes. Furthermore, the method also shows strong applicability in short-term time series datasets. The source code will be made available at <span><span>https://github.com/95xiaoli/FBCD</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"146 ","pages":"Article 105008"},"PeriodicalIF":8.6,"publicationDate":"2025-12-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145659101","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