Egyptian Journal of Remote Sensing and Space Sciences最新文献_第3页

Integrating machine learning with multitemporal remote sensing to quantify spatial soil salinity 结合机器学习与多时相遥感的空间土壤盐分定量研究

IF 4.1 3区地球科学 Q2 ENVIRONMENTAL SCIENCES

Egyptian Journal of Remote Sensing and Space Sciences

Pub Date : 2025-09-01 DOI: 10.1016/j.ejrs.2025.08.005

Rana Muhammad Amir Latif , Adnan Arshad , Jinliao He , Muhammad Habib Ur-Rahman , Fatma Mansour , Ayman El Sabagh , Ibrahim Al-Ashkar

Soil salinization poses a major threat to global agricultural productivity, degrading over 1.5 billion hectares of farmland worldwide. In Pakistan alone, approximately 5.7 million hectares of arable land nearly 30 % of the country’s irrigated area are affected by salinity, leading to substantial crop yield losses. Here, we demonstrate the potential of integrating Remote Sensing (RS) and Machine Learning (ML) to map soil salinity precisely. Using Sentinel-2A and Landsat-8 OLI data, combined with ground measurements of Electrical Conductivity (EC), we trained and validated three ML algorithms: Random Forest (RF), Classification and Regression Tree (CART), and Support Vector Regression (SVR). Through a refined selection process, we identified SI1, SI4, SI5, CRSI, and wetness as the most relevant indicators for soil salinity mapping. Our results show that RF outperforms CART and SVR, achieving R² values of 0.91 (Sentinel-2A) and 0.86 (Landsat-8). The RF maps accurately depicted salt-affected lands, including the Indus River, swamp areas, agricultural fields, and saltpan areas. We estimate that 179,200 ha (Landsat-8) to 207,300 ha (Sentinel-2A) are affected by salinity. This study highlights the applications and integrations of RS and ML for monitoring soil salinity, providing location-specific real-time information for assessing unproductive land and to develop smart management practices and strategies for effective decision making.

土壤盐碱化对全球农业生产力构成重大威胁，导致全球超过15亿公顷农田退化。仅在巴基斯坦，就有大约570万公顷可耕地（约占该国灌溉面积的30%）受到盐碱化影响，导致大量作物减产。在这里，我们展示了整合遥感（RS）和机器学习（ML）来精确绘制土壤盐度的潜力。利用Sentinel-2A和Landsat-8 OLI数据，结合电导率（EC）的地面测量，我们训练并验证了三种机器学习算法：随机森林（RF）、分类与回归树（CART）和支持向量回归（SVR）。通过精细的筛选过程，我们确定SI1、SI4、SI5、CRSI和湿度是与土壤盐度制图最相关的指标。我们的研究结果表明，RF优于CART和SVR， R2值分别为0.91 （Sentinel-2A）和0.86 （Landsat-8）。RF地图准确地描绘了受盐影响的土地，包括印度河、沼泽地区、农田和盐田地区。我们估计179,200公顷（Landsat-8）至207,300公顷（Sentinel-2A）受到盐度的影响。本研究强调了RS和ML在监测土壤盐度方面的应用和集成，为评估非生产性土地提供特定位置的实时信息，并为有效决策制定智能管理实践和策略。

{"title":"Integrating machine learning with multitemporal remote sensing to quantify spatial soil salinity","authors":"Rana Muhammad Amir Latif , Adnan Arshad , Jinliao He , Muhammad Habib Ur-Rahman , Fatma Mansour , Ayman El Sabagh , Ibrahim Al-Ashkar","doi":"10.1016/j.ejrs.2025.08.005","DOIUrl":"10.1016/j.ejrs.2025.08.005","url":null,"abstract":"<div><div>Soil salinization poses a major threat to global agricultural productivity, degrading over 1.5 billion hectares of farmland worldwide. In Pakistan alone, approximately 5.7 million hectares of arable land nearly 30 % of the country’s irrigated area are affected by salinity, leading to substantial crop yield losses. Here, we demonstrate the potential of integrating Remote Sensing (RS) and Machine Learning (ML) to map soil salinity precisely. Using Sentinel-2A and Landsat-8 OLI data, combined with ground measurements of Electrical Conductivity (EC), we trained and validated three ML algorithms: Random Forest (RF), Classification and Regression Tree (CART), and Support Vector Regression (SVR). Through a refined selection process, we identified SI1, SI4, SI5, CRSI, and wetness as the most relevant indicators for soil salinity mapping. Our results show that RF outperforms CART and SVR, achieving R<sup>2</sup> values of 0.91 (Sentinel-2A) and 0.86 (Landsat-8). The RF maps accurately depicted salt-affected lands, including the Indus River, swamp areas, agricultural fields, and saltpan areas. We estimate that 179,200 ha (Landsat-8) to 207,300 ha (Sentinel-2A) are affected by salinity. This study highlights the applications and integrations of RS and ML for monitoring soil salinity, providing location-specific real-time information for assessing unproductive land and to develop smart management practices and strategies for effective decision making.</div></div>","PeriodicalId":48539,"journal":{"name":"Egyptian Journal of Remote Sensing and Space Sciences","volume":"28 3","pages":"Pages 573-586"},"PeriodicalIF":4.1,"publicationDate":"2025-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145004111","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Characterizing the spatiotemporal of near surface seawater intrusion and its association with mangrove distribution in east Lampung-Indonesia 印尼南榜岛东部近地表海水入侵时空特征及其与红树林分布的关系

IF 4.1 3区地球科学 Q2 ENVIRONMENTAL SCIENCES

Egyptian Journal of Remote Sensing and Space Sciences

Pub Date : 2025-08-28 DOI: 10.1016/j.ejrs.2025.08.004

Mochamad Firman Ghazali , Ketut Wikantika , Asep Saepuloh

Numerous research studies have discussed the significance of mangrove forest coverage in protecting coastal regions from seawater intrusion (SWI). Nonetheless, its coverage does not encompass the entire coastline, resulting in the variability of SWI impacts across different coastal regions. This study aims to generate the actual mangrove forest coverage, its change, and the SWI prediction. Therefore, we can quantitatively examine the correlation between changes in mangrove cover and SWI. The Sentinel 2 Multispectral Imaging (MSI), Landsat 5 Thematic Mapper (TM), Landsat 7 Enhanced Thematic Mapper Plus (ETM + ), and Landsat 8 Operational Land Imager and Thermal Infrared Sensor (OLI-TIRS) offered brief and extended analysis durations from 2018 to 2022 and 1990 to 2023, respectively. The SWI is determined by integrating the salt concentration observed in soil and water, using linear and multiple regression models, and the random forest for the near-surface salinity index (NSSI). The predictors for effectively detecting surface salt concentration include the green, red, near-infrared (NIR), and shortwave infrared (SWIR) bands, as well as spectral indices such as the vegetation soil salinity index (VSSI), normalized difference salinity index (NDSI), normalized difference vegetation index (NDVI), and normalized difference water index (NDWI). The findings indicated a linear relationship between SWI and mangrove forest change, since the salinity increases by approximately 5 to 10 % for every hectare of mangrove forest lost. A comprehensive observation from 1990 to 2023 proved that the mangrove forest reduced salinity levels from high to moderate and low, achieving around 49 %, as indicated by a correlation coefficient (R²) of 0.82. This study agreed that mangroves naturally control the SWI in coastal areas.

许多研究都讨论了红树林覆盖率在保护沿海地区免受海水入侵（SWI）方面的意义。然而，其覆盖范围并不包括整个海岸线，导致SWI影响在不同沿海地区的变化。本研究旨在生成红树林的实际覆盖率及其变化，并对SWI进行预测。因此，我们可以定量地考察红树林覆盖变化与SWI的相关性。哨兵2号多光谱成像仪（MSI）、Landsat 5号专题成像仪（TM）、Landsat 7号增强型专题成像仪Plus （ETM +）和Landsat 8号作战陆地成像仪和热红外传感器（OLI-TIRS）分别提供了2018年至2022年和1990年至2023年的短暂和延长的分析时间。SWI是通过综合土壤和水中观测到的盐浓度，使用线性和多元回归模型，以及近地表盐度指数（NSSI）的随机森林来确定的。有效探测地表盐浓度的预测因子包括绿色、红色、近红外（NIR）和短波红外（SWIR）波段，以及植被土壤盐度指数（VSSI）、归一化差异盐度指数（NDSI）、归一化差异植被指数（NDVI）和归一化差异水分指数（NDWI）等光谱指标。研究结果表明SWI与红树林变化之间存在线性关系，因为每公顷红树林消失，盐度增加约5%至10%。1990 - 2023年的综合观测表明，红树林的盐度水平由高到中、低依次降低，相关系数（R2）为0.82，降低幅度约为49%。本研究同意红树林自然控制沿海地区SWI。

{"title":"Characterizing the spatiotemporal of near surface seawater intrusion and its association with mangrove distribution in east Lampung-Indonesia","authors":"Mochamad Firman Ghazali , Ketut Wikantika , Asep Saepuloh","doi":"10.1016/j.ejrs.2025.08.004","DOIUrl":"10.1016/j.ejrs.2025.08.004","url":null,"abstract":"<div><div>Numerous research studies have discussed the significance of mangrove forest coverage in protecting coastal regions from seawater intrusion (SWI). Nonetheless, its coverage does not encompass the entire coastline, resulting in the variability of SWI impacts across different coastal regions. This study aims to generate the actual mangrove forest coverage, its change, and the SWI prediction. Therefore, we can quantitatively examine the correlation between changes in mangrove cover and SWI. The Sentinel 2 Multispectral Imaging (MSI), Landsat 5 Thematic Mapper (TM), Landsat 7 Enhanced Thematic Mapper Plus (ETM + ), and Landsat 8 Operational Land Imager and Thermal Infrared Sensor (OLI-TIRS) offered brief and extended analysis durations from 2018 to 2022 and 1990 to 2023, respectively. The SWI is determined by integrating the salt concentration observed in soil and water, using linear and multiple regression models, and the random forest for the near-surface salinity index (NSSI). The predictors for effectively detecting surface salt concentration include the green, red, near-infrared (NIR), and shortwave infrared (SWIR) bands, as well as spectral indices such as the vegetation soil salinity index (VSSI), normalized difference salinity index (NDSI), normalized difference vegetation index (NDVI), and normalized difference water index (NDWI). The findings indicated a linear relationship between SWI and mangrove forest change, since the salinity increases by approximately 5 to 10 % for every hectare of mangrove forest lost. A comprehensive observation from 1990 to 2023 proved that the mangrove forest reduced salinity levels from high to moderate and low, achieving around 49 %, as indicated by a correlation coefficient (R<sup>2</sup>) of 0.82. This study agreed that mangroves naturally control the SWI in coastal areas.</div></div>","PeriodicalId":48539,"journal":{"name":"Egyptian Journal of Remote Sensing and Space Sciences","volume":"28 3","pages":"Pages 562-572"},"PeriodicalIF":4.1,"publicationDate":"2025-08-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144907268","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Modeling and evaluation of Mardin groundwater level potential using the TOPSIS method 利用TOPSIS方法对Mardin地下水位势进行建模与评价

IF 4.1 3区地球科学 Q2 ENVIRONMENTAL SCIENCES

Egyptian Journal of Remote Sensing and Space Sciences

Pub Date : 2025-08-20 DOI: 10.1016/j.ejrs.2025.08.003

Veysel Aslan, Mehmet Yasar Sepetcioglu

Groundwater is one of the most important resources that can be put into operation in a short time, does not require purification, and has been used to meet drinking, utility, irrigation, and industrial water needs from past to present. Therefore, certain methods must be applied to prevent the degradation of this valuable resource. In other words, the choice of method for groundwater resources is a crucial decision. Selecting a method or methods suitable for the intended work is extremely important for both the future sustainability of groundwater potential and the efficiency of the water to be utilized. In this study, the evaluation of the groundwater potential of Mardin, Turkey, is discussed. To assess groundwater potential, TOPSIS ranking techniques, which are among the Multi-Criteria Decision-Making (MCDM) methods, were utilized. Raster thematic maps of factors such as precipitation, soil type, slope, land use/land cover, geology, drainage density, and geomorphology for the study area were produced using Geographic Information System (GIS) software. Subsequently, classified maps of these produced maps were created. Following this process, a Groundwater Potential Index (GWPI) map was generated based on the values obtained through the application of the TOPSIS method. Considering the effects of the parameters on groundwater, the criteria were weighted with randomly assigned values. In the final stage, the most suitable site selection for the study area was determined using the TOPSIS ranking methods.

地下水是可以在短时间内投入使用的最重要的资源之一，不需要净化，从过去到现在一直用于满足饮用，公用事业，灌溉和工业用水需求。因此，必须采用某些方法来防止这种宝贵资源的退化。换句话说，地下水资源开采方法的选择是一个至关重要的决定。选择一种或几种适合预期工作的方法对于地下水潜力的未来可持续性和所利用的水的效率都是极其重要的。本文讨论了土耳其马尔丁地区地下水潜力的评价。为了评价地下水潜力，采用了多准则决策（MCDM）方法中的TOPSIS排序技术。利用地理信息系统（GIS）软件制作了研究区域降水、土壤类型、坡度、土地利用/土地覆盖、地质、排水密度和地貌等因素的栅格专题图。随后，对这些生成的地图进行分类绘制。在此过程中，根据TOPSIS方法获得的数值生成地下水潜力指数（GWPI）图。考虑到各参数对地下水的影响，采用随机赋值对各指标进行加权。在最后阶段，使用TOPSIS排序方法确定最适合研究区域的选址。

{"title":"Modeling and evaluation of Mardin groundwater level potential using the TOPSIS method","authors":"Veysel Aslan, Mehmet Yasar Sepetcioglu","doi":"10.1016/j.ejrs.2025.08.003","DOIUrl":"10.1016/j.ejrs.2025.08.003","url":null,"abstract":"<div><div>Groundwater is one of the most important resources that can be put into operation in a short time, does not require purification, and has been used to meet drinking, utility, irrigation, and industrial water needs from past to present. Therefore, certain methods must be applied to prevent the degradation of this valuable resource. In other words, the choice of method for groundwater resources is a crucial decision. Selecting a method or methods suitable for the intended work is extremely important for both the future sustainability of groundwater potential and the efficiency of the water to be utilized. In this study, the evaluation of the groundwater potential of Mardin, Turkey, is discussed. To assess groundwater potential, TOPSIS ranking techniques, which are among the Multi-Criteria Decision-Making (MCDM) methods, were utilized. Raster thematic maps of factors such as precipitation, soil type, slope, land use/land cover, geology, drainage density, and geomorphology for the study area were produced using Geographic Information System (GIS) software. Subsequently, classified maps of these produced maps were created. Following this process, a Groundwater Potential Index (GWPI) map was generated based on the values obtained through the application of the TOPSIS method. Considering the effects of the parameters on groundwater, the criteria were weighted with randomly assigned values. In the final stage, the most suitable site selection for the study area was determined using the TOPSIS ranking methods.</div></div>","PeriodicalId":48539,"journal":{"name":"Egyptian Journal of Remote Sensing and Space Sciences","volume":"28 3","pages":"Pages 553-561"},"PeriodicalIF":4.1,"publicationDate":"2025-08-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144866186","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Mapping seasonal soil deformation in expansive clay using synthetic aperture radar interferometry: A case study in Diamniadio, Senegal 利用合成孔径雷达干涉测量法测绘膨胀粘土的季节性土壤变形：以塞内加尔Diamniadio为例

IF 4.1 3区地球科学 Q2 ENVIRONMENTAL SCIENCES

Egyptian Journal of Remote Sensing and Space Sciences

Pub Date : 2025-08-12 DOI: 10.1016/j.ejrs.2025.08.002

Seynabou Toure , Oluwaseyi Dasho , Souleye Wade , Oumar Diop , Kidiyo Kpalma , Amadou S. Maiga

Expansive clay soils, known for their moisture-driven volumetric changes, pose a critical challenge to infrastructure stability in rapidly urbanizing regions. This study presents the first quantitative assessment of seasonal soil deformation in Diamniadio, Senegal, using Persistent Scatterer and Distributed Scatterer Interferometric Synthetic Aperture Radar (PSDS-InSAR) techniques with Sentinel-1 data from March 2017 to July 2024. High-resolution time series and deformation maps were generated for 11 strategic urban sites using 793 interferograms processed via a wavelet-based InSAR approach. Results reveal a clear pattern of seasonal uplift during the wet season (July–October) and subsidence during the dry season (October–June), with vertical deformation amplitudes ranging from 0.5 to 5 mm. Localized subsidence was detected in key areas such as the United Nations House (−16.11 mm/year) and Dakar Arena (−2.28 mm/year), correlating with active construction and soil sensitivity. Angular distortion analysis identified critical zones where differential settlement exceeds empirical thresholds for structural damage, with total angular distortion values reaching up to 2.5 × 10⁻³. An exposure map combining deformation gradients and infrastructure distribution highlights high-risk zones, particularly in clay-rich soil areas. These findings provide a robust spatial and temporal characterization of soil behavior, offering essential insights for geotechnical hazard assessment and sustainable urban development in Diamniadio and similar contexts.

膨胀粘土以其水分驱动的体积变化而闻名，对快速城市化地区的基础设施稳定性构成了重大挑战。利用Sentinel-1 2017年3月至2024年7月的持续散射体和分布式散射体干涉合成孔径雷达（PSDS-InSAR）技术，首次对塞内加尔Diamniadio地区的季节性土壤变形进行了定量评估。利用基于小波的InSAR方法处理的793张干涉图，生成了11个战略城市站点的高分辨率时间序列和变形图。研究结果表明，黄土高原在湿季（7 ~ 10月）表现出明显的季节性抬升和旱季（10 ~ 6月）沉降的特征，垂直变形幅度在0.5 ~ 5 mm之间。在联合国大厦（- 16.11 mm/年）和达喀尔体育馆（- 2.28 mm/年）等关键区域检测到局部沉降，这与主动施工和土壤敏感性有关。角畸变分析确定了差异沉降超过经验阈值的关键区域，总角畸变值高达2.5 × 10−3。结合变形梯度和基础设施分布的暴露图突出了高风险区域，特别是在富含粘土的土壤地区。这些发现为土壤行为提供了强有力的时空特征，为Diamniadio和类似环境下的岩土灾害评估和可持续城市发展提供了重要见解。

{"title":"Mapping seasonal soil deformation in expansive clay using synthetic aperture radar interferometry: A case study in Diamniadio, Senegal","authors":"Seynabou Toure , Oluwaseyi Dasho , Souleye Wade , Oumar Diop , Kidiyo Kpalma , Amadou S. Maiga","doi":"10.1016/j.ejrs.2025.08.002","DOIUrl":"10.1016/j.ejrs.2025.08.002","url":null,"abstract":"<div><div>Expansive clay soils, known for their moisture-driven volumetric changes, pose a critical challenge to infrastructure stability in rapidly urbanizing regions. This study presents the first quantitative assessment of seasonal soil deformation in Diamniadio, Senegal, using Persistent Scatterer and Distributed Scatterer Interferometric Synthetic Aperture Radar (PSDS-InSAR) techniques with Sentinel-1 data from March 2017 to July 2024. High-resolution time series and deformation maps were generated for 11 strategic urban sites using 793 interferograms processed via a wavelet-based InSAR approach. Results reveal a clear pattern of seasonal uplift during the wet season (July–October) and subsidence during the dry season (October–June), with vertical deformation amplitudes ranging from 0.5 to 5 mm. Localized subsidence was detected in key areas such as the United Nations House (−16.11 mm/year) and Dakar Arena (−2.28 mm/year), correlating with active construction and soil sensitivity. Angular distortion analysis identified critical zones where differential settlement exceeds empirical thresholds for structural damage, with total angular distortion values reaching up to 2.5 × 10<sup>−3</sup>. An exposure map combining deformation gradients and infrastructure distribution highlights high-risk zones, particularly in clay-rich soil areas. These findings provide a robust spatial and temporal characterization of soil behavior, offering essential insights for geotechnical hazard assessment and sustainable urban development in Diamniadio and similar contexts.</div></div>","PeriodicalId":48539,"journal":{"name":"Egyptian Journal of Remote Sensing and Space Sciences","volume":"28 3","pages":"Pages 542-552"},"PeriodicalIF":4.1,"publicationDate":"2025-08-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144813945","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

TriM-Net: Trinityformer-Mamba fusion for road extraction in remote sensing trimnet：用于遥感道路提取的Trinityformer-Mamba融合

IF 4.1 3区地球科学 Q2 ENVIRONMENTAL SCIENCES

Egyptian Journal of Remote Sensing and Space Sciences

Pub Date : 2025-08-05 DOI: 10.1016/j.ejrs.2025.07.006

Zhenzhong Huang , Hongjuan Shao , Chao Ren , Hongman Li , Haoming Bai , Zhou Lei , Gu Yao , Qinyi Chen

Precise road information extraction is crucial for transportation and intelligent sensing. Recently, the fusion of CNN and Transformer architectures in remote sensing-based road extraction, along with U-shaped semantic segmentation networks, has gained significant attention. However, existing methods rely heavily on global features while overlooking local details, limiting accuracy in complex road scenarios. To address this, we propose Trinityformer-Mamba Network (TriM-Net) to enhance local feature extraction. TriM-Net adopts Trinityformer, a modified Transformer architecture. This architecture optimizes local feature perception and reduces computational overhead by replacing the traditional softmax with an improved self-attention mechanism and a novel normalization method. The feedforward network employs a Kolmogorov-Arnold network (KAN), reducing neuron count while enhancing local detail capture using edge activation functions and the Arnold transform. Additionally, the normalization layer integrates the benefits of BatchNorm and LayerNorm for better performance. Furthermore, TriM-Net incorporates an MT_block built with stacked Mamba networks. By leveraging their internal CausalConv1D and SSM modules, this block enhances modeling and local perception while effectively merging Transformer and CNN information for improved image reconstruction. Experimental results demonstrate TriM-Net’s significant superiority over existing state-of-the-art models. On the LSRV dataset, it outperformed the second-best model with advantages of 2.17% in Precision, 0.34% in Recall, 1.72% in IoU, and 2.09% in F1-score. Similarly, on the Massachusetts Road Dataset, it achieved superior Recall (0.45%), IoU (1.41%), and F1-score (1.07%) over its closest competitor. These substantial improvements highlight TriM-Net’s outstanding performance in road information extraction.

精确的道路信息提取对交通运输和智能传感至关重要。近年来，CNN和Transformer架构的融合以及u型语义分割网络在基于遥感的道路提取中得到了广泛关注。然而，现有的方法严重依赖全局特征，而忽略了局部细节，限制了复杂道路场景的准确性。为了解决这个问题，我们提出了Trinityformer-Mamba Network （TriM-Net）来增强局部特征提取。TriM-Net采用了一种改进的Transformer架构Trinityformer。该体系结构通过改进的自关注机制和新的归一化方法取代传统的softmax，优化了局部特征感知，减少了计算开销。前馈网络采用了Kolmogorov-Arnold网络（KAN），减少了神经元数量，同时利用边缘激活函数和Arnold变换增强了局部细节捕获。此外，规范化层集成了BatchNorm和LayerNorm的优点，以获得更好的性能。此外，TriM-Net还集成了一个MT_block，该MT_block由堆叠的Mamba网络构建。通过利用其内部的CausalConv1D和SSM模块，该块增强了建模和局部感知，同时有效地合并Transformer和CNN信息，以改进图像重建。实验结果表明，TriM-Net比现有的最先进模型具有显著的优势。在LSRV数据集上，它以2.17%的精度、0.34%的召回率、1.72%的IoU和2.09%的F1-score优势优于次优模型。同样，在马萨诸塞州道路数据集上，它比最接近的竞争对手取得了更高的召回率（0.45%）、IoU（1.41%）和f1分数（1.07%）。这些实质性的改进凸显了TriM-Net在道路信息提取方面的卓越性能。

{"title":"TriM-Net: Trinityformer-Mamba fusion for road extraction in remote sensing","authors":"Zhenzhong Huang , Hongjuan Shao , Chao Ren , Hongman Li , Haoming Bai , Zhou Lei , Gu Yao , Qinyi Chen","doi":"10.1016/j.ejrs.2025.07.006","DOIUrl":"10.1016/j.ejrs.2025.07.006","url":null,"abstract":"<div><div>Precise road information extraction is crucial for transportation and intelligent sensing. Recently, the fusion of CNN and Transformer architectures in remote sensing-based road extraction, along with U-shaped semantic segmentation networks, has gained significant attention. However, existing methods rely heavily on global features while overlooking local details, limiting accuracy in complex road scenarios. To address this, we propose Trinityformer-Mamba Network (TriM-Net) to enhance local feature extraction. TriM-Net adopts Trinityformer, a modified Transformer architecture. This architecture optimizes local feature perception and reduces computational overhead by replacing the traditional softmax with an improved self-attention mechanism and a novel normalization method. The feedforward network employs a Kolmogorov-Arnold network (KAN), reducing neuron count while enhancing local detail capture using edge activation functions and the Arnold transform. Additionally, the normalization layer integrates the benefits of BatchNorm and LayerNorm for better performance. Furthermore, TriM-Net incorporates an MT_block built with stacked Mamba networks. By leveraging their internal CausalConv1D and SSM modules, this block enhances modeling and local perception while effectively merging Transformer and CNN information for improved image reconstruction. Experimental results demonstrate TriM-Net’s significant superiority over existing state-of-the-art models. On the LSRV dataset, it outperformed the second-best model with advantages of 2.17% in Precision, 0.34% in Recall, 1.72% in IoU, and 2.09% in F1-score. Similarly, on the Massachusetts Road Dataset, it achieved superior Recall (0.45%), IoU (1.41%), and F1-score (1.07%) over its closest competitor. These substantial improvements highlight TriM-Net’s outstanding performance in road information extraction.</div></div>","PeriodicalId":48539,"journal":{"name":"Egyptian Journal of Remote Sensing and Space Sciences","volume":"28 3","pages":"Pages 523-533"},"PeriodicalIF":4.1,"publicationDate":"2025-08-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144780611","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Evaluating the factors affecting landslides using machine learning algorithms (case study: the catchment area of Karun-3 Dam, Iran) 利用机器学习算法评估影响滑坡的因素（案例研究：伊朗Karun-3大坝集水区）

IF 4.1 3区地球科学 Q2 ENVIRONMENTAL SCIENCES

Egyptian Journal of Remote Sensing and Space Sciences

Pub Date : 2025-07-31 DOI: 10.1016/j.ejrs.2025.07.005

Rahman Zandi , Ghasem Shah Pari Far

Landslides are among the phenomena associated with environmental impacts and human and financial losses worldwide. Investigating environmental issues such as landslides and preparing hazard maps are essential for managers and planners. This study examines and models landslides in the catchment area of Karun-3 Dam located in Khuzestan province, Iran, using six machine learning algorithms, including Random Forest (RF), Boosted Regression Tree (BRT), Generalized Aggregate Model (GAM), Support Vector Model (SVM), Classification and Regression Tree (CART), and Generalized Linear Model (GLM). Thirteen independent parameters were identified as the main parameters. Then, their correlation and effects were examined using 284 old landslides, and machine learning models were validated using efficiency, sensitivity, and accuracy indicators. The validation results showed that although all the models used have sufficient accuracy, the RF model (AUC = 0.982, Efficiency = 0.943) has more accuracy than the other five models. Also, the impact of different factors on landslide generation in various models is not the same. In general, the significance of the mentioned parameters is in the range of 0.043 and 0.160. Comparing the results of different models using a non-parametric test shows more similarities between the models used. In general, the results of various models show that the risk of landslides is generally higher on the steep banks of rivers, in the vicinity of lakes, dams, and roads, and especially in lands with soft lithology such as marl. This fact shows us the influence of anthropogenic factors and natural factors simultaneously.

山体滑坡是全球范围内与环境影响、人员和经济损失相关的现象之一。对管理人员和规划人员来说，调查诸如滑坡之类的环境问题和编制灾害地图是必不可少的。本研究利用随机森林（RF）、增强回归树（BRT）、广义聚合模型（GAM）、支持向量模型（SVM）、分类与回归树（CART）和广义线性模型（GLM）等六种机器学习算法，对伊朗胡齐斯坦省Karun-3大坝集水区的滑坡进行了研究和建模。确定了13个独立参数作为主要参数。然后，使用284个老滑坡来检验它们的相关性和影响，并使用效率、灵敏度和准确性指标验证机器学习模型。验证结果表明，虽然所使用的所有模型都具有足够的准确性，但RF模型（AUC = 0.982, Efficiency = 0.943）的准确性高于其他5种模型。不同因素对不同模型滑坡生成的影响也不尽相同。总的来说，上述参数的显著性在0.043 ~ 0.160之间。使用非参数检验比较不同模型的结果显示所使用的模型之间有更多的相似性。总的来说，各种模型的结果表明，在陡峭的河岸、湖泊、水坝和道路附近，特别是在泥沼等岩性较软的土地上，发生山体滑坡的风险通常较高。这一事实同时向我们表明了人为因素和自然因素的影响。

{"title":"Evaluating the factors affecting landslides using machine learning algorithms (case study: the catchment area of Karun-3 Dam, Iran)","authors":"Rahman Zandi , Ghasem Shah Pari Far","doi":"10.1016/j.ejrs.2025.07.005","DOIUrl":"10.1016/j.ejrs.2025.07.005","url":null,"abstract":"<div><div>Landslides are among the phenomena associated with environmental impacts and human and financial losses worldwide. Investigating environmental issues such as landslides and preparing hazard maps are essential for managers and planners. This study examines and models landslides in the catchment area of Karun-3 Dam located in Khuzestan province, Iran, using six machine learning algorithms, including Random Forest (RF), Boosted Regression Tree (BRT), Generalized Aggregate Model (GAM), Support Vector Model (SVM), Classification and Regression Tree (CART), and Generalized Linear Model (GLM). Thirteen independent parameters were identified as the main parameters. Then, their correlation and effects were examined using 284 old landslides, and machine learning models were validated using efficiency, sensitivity, and accuracy indicators. The validation results showed that although all the models used have sufficient accuracy, the RF model (AUC = 0.982, Efficiency = 0.943) has more accuracy than the other five models. Also, the impact of different factors on landslide generation in various models is not the same. In general, the significance of the mentioned parameters is in the range of 0.043 and 0.160. Comparing the results of different models using a non-parametric test shows more similarities between the models used. In general, the results of various models show that the risk of landslides is generally higher on the steep banks of rivers, in the vicinity of lakes, dams, and roads, and especially in lands with soft lithology such as marl. This fact shows us the influence of anthropogenic factors and natural factors simultaneously.</div></div>","PeriodicalId":48539,"journal":{"name":"Egyptian Journal of Remote Sensing and Space Sciences","volume":"28 3","pages":"Pages 512-522"},"PeriodicalIF":4.1,"publicationDate":"2025-07-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144739062","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Advancements and applications of space borne of remote sensing in climate change research: A scoping review 空间遥感在气候变化研究中的进展与应用综述

IF 3.7 3区地球科学 Q2 ENVIRONMENTAL SCIENCES

Egyptian Journal of Remote Sensing and Space Sciences

Pub Date : 2025-07-23 DOI: 10.1016/j.ejrs.2025.07.004

Ricky Anak Kemarau , Zaini Sakawi , Khairul Nizam Abdul Maulud , Wan Shafrina Wan Mohd Jaafar , Stanley Anak Suab , Oliver Valentine Eboy , Nik Norliati Fitri Md Nor , Zulfaqar Sa’adi

This scoping review explores the progress and applications of space-borne remote sensing within the realm of climate change research. It systematically compiles significant advancements in remote sensing technology, with a focus on its application for tracking diverse indicators of climate change. The review performs a thorough examination of various sensor types and technologies, evaluates the challenges and limitations encountered, and considers methods to overcome these obstacles. By adopting an integrated and multidisciplinary approach, the study connects the gap between technological progress and its policy implications, alongside mitigation and adaptation strategies. This offers a holistic view of the pivotal role of remote sensing in the study of climate change, providing valuable insights for researchers, policymakers, and practitioners alike.

本文综述了星载遥感在气候变化研究领域的进展和应用。它系统地汇编了遥感技术方面的重大进展，重点是遥感技术在跟踪各种气候变化指标方面的应用。该综述对各种传感器类型和技术进行了彻底的检查，评估了遇到的挑战和限制，并考虑了克服这些障碍的方法。通过采用综合的多学科方法，该研究将技术进步与其政策影响之间的差距与缓解和适应战略联系起来。这为遥感在气候变化研究中的关键作用提供了一个整体的观点，为研究人员、政策制定者和实践者提供了有价值的见解。

引用次数: 0

Assessment of the hydrogeological potential of the north-eastern sector of the town of Dschang (West Cameroon) using integrated remote sensing, geophysics and multi-criteria analysis 利用综合遥感、地球物理和多标准分析评估Dschang镇（喀麦隆西部）东北地区的水文地质潜力

IF 3.7 3区地球科学 Q2 ENVIRONMENTAL SCIENCES

Egyptian Journal of Remote Sensing and Space Sciences

Pub Date : 2025-07-23 DOI: 10.1016/j.ejrs.2025.07.003

Kenfack Jean Victor, Talla Toteu Rodrigue, Bomeni Isaac Yannick, Demanou Messe Malick Rosvelt, Tchomtchoua Tagne Stéphane, Djoumete Kengni Annie Christelle, Kengni Lucas

This study focuses on the hydrogeological mapping of Dschang, western Cameroon, where drinking water shortages persist due to limited understanding of local aquifers. The research integrates remote sensing, geophysics, and multi-criteria analysis to assess groundwater potential. Key findings include the identification of a primary fracturing network (directions N 20°–30°E and N 60°–70°E) and three distinct resistivity domains based on vertical electrical soundings carried out on 120 points. The resistivity values range from 1.43 to 2467.429 Ω.m, classified as conductive, less conductive, or resistant domains. Hydraulic parameters such as conductivity (0.0036–116.0073 m/day), porosity (0.192–46.894 %), transmissivity (0.019–1507.817 m²/day), and aquifer thickness (2–63 m) were analyzed. Using multi-criteria analysis, the data were synthesized to produce a hydrogeological map. Highly favorable zones for groundwater exploitation are concentrated in basaltic and ignimbritic formations in the north and south of the study area, while moderately favorable zones surround these areas. Unfavorable zones are located in the center and southern periphery.

这项研究的重点是喀麦隆西部Dschang的水文地质测绘，由于对当地含水层的了解有限，那里的饮用水短缺问题仍然存在。该研究综合了遥感、地球物理和多准则分析来评估地下水潜力。主要发现包括确定了一个主压裂网络（北纬20°-30°E和北纬60°-70°E），以及基于在120个点上进行的垂直电测深的三个不同的电阻率域。电阻率取值范围为1.43 ~ 2467.429 Ω。M，分为导电，不导电或电阻域。分析了含水层电导率（0.0036-116.0073 m/day）、孔隙度（0.192-46.894 %）、透光率（0.019-1507.817 m2/day）、含水层厚度（2-63 m）等水力参数。采用多准则分析方法，对资料进行综合处理，生成水文地质图。研究区北部和南部的玄武岩组和火成岩组为地下水开发的高度有利带，而这些区域的周围为中等有利带。不利区域位于中心和南部边缘。

{"title":"Assessment of the hydrogeological potential of the north-eastern sector of the town of Dschang (West Cameroon) using integrated remote sensing, geophysics and multi-criteria analysis","authors":"Kenfack Jean Victor, Talla Toteu Rodrigue, Bomeni Isaac Yannick, Demanou Messe Malick Rosvelt, Tchomtchoua Tagne Stéphane, Djoumete Kengni Annie Christelle, Kengni Lucas","doi":"10.1016/j.ejrs.2025.07.003","DOIUrl":"10.1016/j.ejrs.2025.07.003","url":null,"abstract":"<div><div>This study focuses on the hydrogeological mapping of Dschang, western Cameroon, where drinking water shortages persist due to limited understanding of local aquifers. The research integrates remote sensing, geophysics, and multi-criteria analysis to assess groundwater potential. Key findings include the identification of a primary fracturing network (directions N 20°–30°E and N 60°–70°E) and three distinct resistivity domains based on vertical electrical soundings carried out on 120 points. The resistivity values range from 1.43 to 2467.429 Ω.m, classified as conductive, less conductive, or resistant domains. Hydraulic parameters such as conductivity (0.0036–116.0073 m/day), porosity (0.192–46.894 %), transmissivity (0.019–1507.817 m<sup>2</sup>/day), and aquifer thickness (2–63 m) were analyzed. Using multi-criteria analysis, the data were synthesized to produce a hydrogeological map. Highly favorable zones for groundwater exploitation are concentrated in basaltic and ignimbritic formations in the north and south of the study area, while moderately favorable zones surround these areas. Unfavorable zones are located in the center and southern periphery.</div></div>","PeriodicalId":48539,"journal":{"name":"Egyptian Journal of Remote Sensing and Space Sciences","volume":"28 3","pages":"Pages 484-511"},"PeriodicalIF":3.7,"publicationDate":"2025-07-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144685770","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Artificial intelligence enabled spectral-spatial feature extraction techniques for land use and land cover classification using hyperspectral images – An inclusive review 使用高光谱图像进行土地利用和土地覆盖分类的人工智能光谱空间特征提取技术-综述

IF 3.7 3区地球科学 Q2 ENVIRONMENTAL SCIENCES

Egyptian Journal of Remote Sensing and Space Sciences

Pub Date : 2025-07-18 DOI: 10.1016/j.ejrs.2025.06.004

V. Sangeetha, L. Agilandeeswari

The growth of artificial intelligence techniques such as machine learning and deep learning facilitates the hyperspectral image processing applicable in developing various remote sensing applications such as Change detection in Land Use and Land Cover (LULC) classification, Evaluation of the nutritional content, and health of the crops in Agriculture. However, Hyperspectral imaging is frequently utilized in remote sensing and earth observation applications to identify environmental changes. One of the key tasks in hyperspectral image classification is feature extraction. This paper gives a comprehensive review of the recent hyperspectral image feature extraction techniques for LULC. This study aims to identify the open issues, research challenges, and future directions that will help researchers develop efficient feature extraction techniques for better LULC hyperspectral image classification. The performance of the state-of-the-art feature extraction techniques for hyperspectral images is analyzed in terms of the overall accuracy, average accuracy, and kappa coefficient across the benchmark datasets, namely Indian Pines, Pavia dataset, and Salinas dataset. From the analysis, we observe that in all the benchmark datasets, the framework 2D + 3D CNN with spectral-spatial integration not only extracts the comprehensive features but also increases the classification accuracy with less computational complexity compared to other competing frameworks. Both 2D CNNs and 3D CNNs are utilized for extracting features and patterns from data with multiple spectral bands, and each architecture has its advantages and challenges. 2D CNNs are more common and computationally efficient, while 3D CNNs capture spatial-spectral correlations more directly.

机器学习和深度学习等人工智能技术的发展促进了高光谱图像处理，可用于开发各种遥感应用，如土地利用和土地覆盖变化检测（LULC）分类、农业作物营养成分评估和健康状况评估。然而，高光谱成像在遥感和地球观测应用中经常被用于识别环境变化。高光谱图像分类的关键任务之一是特征提取。本文综述了近年来用于LULC的高光谱图像特征提取技术。本研究旨在确定开放的问题、研究挑战和未来的方向，这将有助于研究人员开发有效的特征提取技术，以更好地进行LULC高光谱图像分类。从总体精度、平均精度和kappa系数三个方面分析了目前最先进的高光谱图像特征提取技术在基准数据集（即Indian Pines、Pavia和Salinas数据集）中的性能。通过分析，我们发现在所有的基准数据集中，与其他竞争框架相比，具有光谱-空间集成的2D + 3D CNN框架不仅提取了综合特征，而且在计算复杂度较低的情况下提高了分类精度。二维cnn和三维cnn都用于从多光谱波段的数据中提取特征和模式，每种架构都有其优势和挑战。2D cnn更常见，计算效率更高，而3D cnn更直接地捕获空间-光谱相关性。

{"title":"Artificial intelligence enabled spectral-spatial feature extraction techniques for land use and land cover classification using hyperspectral images – An inclusive review","authors":"V. Sangeetha, L. Agilandeeswari","doi":"10.1016/j.ejrs.2025.06.004","DOIUrl":"10.1016/j.ejrs.2025.06.004","url":null,"abstract":"<div><div>The growth of artificial intelligence techniques such as machine learning and deep learning facilitates the hyperspectral image processing applicable in developing various remote sensing applications such as Change detection in Land Use and Land Cover (LULC) classification, Evaluation of the nutritional content, and health of the crops in Agriculture. However, Hyperspectral imaging is frequently utilized in remote sensing and earth observation applications to identify environmental changes. One of the key tasks in hyperspectral image classification is feature extraction. This paper gives a comprehensive review of the recent hyperspectral image feature extraction techniques for LULC. This study aims to identify the open issues, research challenges, and future directions that will help researchers develop efficient feature extraction techniques for better LULC hyperspectral image classification. The performance of the state-of-the-art feature extraction techniques for hyperspectral images is analyzed in terms of the overall accuracy, average accuracy, and kappa coefficient across the benchmark datasets, namely Indian Pines, Pavia dataset, and Salinas dataset. From the analysis, we observe that in all the benchmark datasets, the framework 2D + 3D CNN with spectral-spatial integration not only extracts the comprehensive features but also increases the classification accuracy with less computational complexity compared to other competing frameworks. Both 2D CNNs and 3D CNNs are utilized for extracting features and patterns from data with multiple spectral bands, and each architecture has its advantages and challenges. 2D CNNs are more common and computationally efficient, while 3D CNNs capture spatial-spectral correlations more directly.</div></div>","PeriodicalId":48539,"journal":{"name":"Egyptian Journal of Remote Sensing and Space Sciences","volume":"28 3","pages":"Pages 455-467"},"PeriodicalIF":3.7,"publicationDate":"2025-07-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144662049","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

A novel weighted average ensemble method for landslide susceptibility mapping: A case study in Yuanyang, China 一种新的加权平均集合方法在滑坡易感性制图中的应用——以元阳为例

IF 3.7 3区地球科学 Q2 ENVIRONMENTAL SCIENCES

Egyptian Journal of Remote Sensing and Space Sciences

Pub Date : 2025-07-14 DOI: 10.1016/j.ejrs.2025.07.002

Valisoasarobidy José Gabriel , Ruihong Wang , Doshrot Mahato , Can Wei

Landslide susceptibility mapping is critical for risk assessment, but existing ensemble methods like VotingClassifier suffer from three unresolved limitations: static weight allocation that ignores spatial variability, lack of quantifiable uncertainty measures, and poor integration of interpretability tools. This study introduces a novel weighted average ensemble method that dynamically adjusts weights for Random Forest (RF), Support Vector Machine (SVM), and Extreme Gradient Boosting (XGBoost) through 5-fold spatial cross-validation, improving prediction robustness across Yuanyang County’s 2240 km² of mountainous terrain (23°05′–23°15′N, 102°40′–102°50′E) with 817 validated landslides. The method tackles important issues by combining the best features of strong models while reducing the effects of related variables using composite indices (like a soil-lithology index based on a Pearson correlation of r = 0.81), backed by a thorough preprocessing process that includes Moran’s I-validated stratified sampling (I = 0.12), normalization that accounts for outliers (95th percentile), and spatial division with 500 m buffers. The novel ensemble achieved an accuracy of 84.32 % and an ROC AUC of 91.96 %, with sensitivity analysis via SHAP (SHapley Additive exPlanations) identifying rainfall (21 %), distance index (13 %), and elevation slope index (27 %) as dominant drivers, while uncertainty analysis revealed prediction intervals of ±0.62 width (95 % coverage). The resulting maps, validated through spatial consistency checks (AUC > 0.84), provide actionable tools for high-risk zones. This research improves landslide susceptibility mapping by developing a dynamic, uncertainty-based system that rectifies major weaknesses in static ensemble methods, thereby establishing a replicable standard for future investigations.

滑坡敏感性制图对于风险评估至关重要，但现有的集成方法（如VotingClassifier）存在三个未解决的限制：忽略空间变异性的静态权重分配，缺乏可量化的不确定性度量，以及对可解释性工具的集成能力差。本文提出了一种新的加权平均集成方法，通过5倍空间交叉验证，动态调整随机森林（RF）、支持向量机（SVM）和极端梯度增强（XGBoost）的权重，提高了对远阳县2240 km2山地地形（23°05′-23°15′n, 102°40′-102°50′e） 817个已验证滑坡的预测鲁棒性。该方法通过结合强模型的最佳特征来解决重要问题，同时使用复合指数（如基于r = 0.81的Pearson相关性的土壤-岩石指数）减少相关变量的影响，并辅以彻底的预处理过程，包括Moran的I验证分层抽样（I = 0.12），考虑异常值的归一化（第95百分位数），以及500 m缓冲区的空间划分。新集合的准确度为84.32%，ROC AUC为91.96%，通过SHapley加性解释（SHapley Additive exPlanations）进行敏感性分析，确定降雨（21%）、距离指数（13%）和高程坡度指数（27%）是主要驱动因素，而不确定性分析显示预测区间为±0.62宽度（95%覆盖率）。生成的地图，通过空间一致性检查(AUC >；0.84)，为高风险地区提供可操作的工具。本研究通过开发一个动态的、基于不确定性的系统来改进滑坡易感性制图，该系统纠正了静态集合方法的主要弱点，从而为未来的调查建立了可复制的标准。

{"title":"A novel weighted average ensemble method for landslide susceptibility mapping: A case study in Yuanyang, China","authors":"Valisoasarobidy José Gabriel , Ruihong Wang , Doshrot Mahato , Can Wei","doi":"10.1016/j.ejrs.2025.07.002","DOIUrl":"10.1016/j.ejrs.2025.07.002","url":null,"abstract":"<div><div>Landslide susceptibility mapping is critical for risk assessment, but existing ensemble methods like VotingClassifier suffer from three unresolved limitations: static weight allocation that ignores spatial variability, lack of quantifiable uncertainty measures, and poor integration of interpretability tools. This study introduces a novel weighted average ensemble method that dynamically adjusts weights for Random Forest (RF), Support Vector Machine (SVM), and Extreme Gradient Boosting (XGBoost) through 5-fold spatial cross-validation, improving prediction robustness across Yuanyang County’s 2240 km<sup>2</sup> of mountainous terrain (23°05′–23°15′N, 102°40′–102°50′E) with 817 validated landslides. The method tackles important issues by combining the best features of strong models while reducing the effects of related variables using composite indices (like a soil-lithology index based on a Pearson correlation of r = 0.81), backed by a thorough preprocessing process that includes Moran’s I-validated stratified sampling (I = 0.12), normalization that accounts for outliers (95th percentile), and spatial division with 500 m buffers. The novel ensemble achieved an accuracy of 84.32 % and an ROC AUC of 91.96 %, with sensitivity analysis via SHAP (SHapley Additive exPlanations) identifying rainfall (21 %), distance index (13 %), and elevation slope index (27 %) as dominant drivers, while uncertainty analysis revealed prediction intervals of ±0.62 width (95 % coverage). The resulting maps, validated through spatial consistency checks (AUC > 0.84), provide actionable tools for high-risk zones. This research improves landslide susceptibility mapping by developing a dynamic, uncertainty-based system that rectifies major weaknesses in static ensemble methods, thereby establishing a replicable standard for future investigations.</div></div>","PeriodicalId":48539,"journal":{"name":"Egyptian Journal of Remote Sensing and Space Sciences","volume":"28 3","pages":"Pages 436-454"},"PeriodicalIF":3.7,"publicationDate":"2025-07-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144614052","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0