International journal of applied earth observation and geoinformation : ITC journal最新文献

{"title":"SFD-YOLO: A novel framework for subsidence funnels detection in China based on large-scale SAR interferograms","authors":"Jing Guo ,&nbsp;Zhengjia Zhang ,&nbsp;Peifeng Ma ,&nbsp;Mengmeng Wang ,&nbsp;Xuefei Zhang ,&nbsp;Dongdong Li ,&nbsp;Bing Sui","doi":"10.1016/j.jag.2025.104605","DOIUrl":"10.1016/j.jag.2025.104605","url":null,"abstract":"<div><div>Accurate identification of subsidence funnels is essential for assessing surface deformation in mining areas, preventing disasters, and optimizing resource management. However, recognizing subsidence funnels of varying sizes in large-scale interferograms poses significant challenges, particularly for small-sized funnels. Their indistinct features and susceptibility to background noise interference often result in suboptimal detection accuracy. To address these challenges, this study proposes a deep learning network based on the YOLO architecture—SFD-YOLO (Sinking Funnel Detection-YOLO). The model incorporates the DWR-C2f module, which enhances multi-scale feature extraction and significantly improves the detection of small-sized subsidence funnels. Additionally, the innovative Inner-WIoU regression loss function improves the localization accuracy of detection boxes while also alleviates the imbalance between hard and easy samples. Experimental results demonstrate that the fully trained SFD-YOLO model achieves an mAP50 accuracy of 92.00% while maintaining high efficiency, significantly outperforming other advanced methods. Applying the SFD-YOLO model to interferograms across China detected a total of 3,842 subsidence funnels, with Shanxi, Inner Mongolia, Shaanxi, and Anhui identified as the four provinces with the highest funnel number. Overall, subsidence funnels are predominantly distributed in northern and northwestern China. Further analysis and experimental evaluation reveal that the SFD-YOLO model exhibits strong generalization capabilities across complex surface environments nationwide and multi-source satellite data.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"140 ","pages":"Article 104605"},"PeriodicalIF":7.6,"publicationDate":"2025-05-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144084539","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}

{"title":"Machine learning and clustering for supporting the identification of active landslides at national scale","authors":"Camilla Medici ,&nbsp;Alessandro Novellino ,&nbsp;Claire Dashwood ,&nbsp;Silvia Bianchini","doi":"10.1016/j.jag.2025.104608","DOIUrl":"10.1016/j.jag.2025.104608","url":null,"abstract":"<div><div>Landslides are one of the major geohazards causing significant economic damage and loss of life, with impacts expected to increase due to climate change. Landslide inventory maps (LIMs) are essential for risk management and reduction, but they usually remain an overlooked issue, especially over very large areas, i.e. at a regional or national level. Nowadays, extensive interferometric satellite radar data with wide area coverage are profitably available, but their potential can be not fully exploited due to the challenge of managing them. In this context, we used space-borne advanced Interferometric Synthetic Aperture Radar (InSAR) data at a national scale, in order to create a useful database of active slope instability movement areas to rely on where the landslide inventory map is missing or largely incomplete. Specifically, we provide insights into a new approach, proposing a national-scale method that combines Machine Learning (ML) and clustering tools, which are crucial to manage a huge amount of data. The proposed methodology has been applied to Great Britain. The use of a ML algorithm, specifically Hierarchical Density-Based Spatial Clustering of Applications with Noise (HDBSCAN) for noise filtering, has allowed the InSAR dataset to be reduced from approximately 6.5 million points to about 3.8 million points per component. Thus, implementing ML along with Slope Units for geomorphological reliability, and tools for identifying and classifying active deformation areas yields an InSAR landslide inventory map. Through this process, 336,557 Slope Units have been classified; of these, 5% show discrepancies between landslide inventory and InSAR data. Identifying these areas, along with those classified as landslides by both datasets, is crucial for risk management as it highlights areas that require closer inspections.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"140 ","pages":"Article 104608"},"PeriodicalIF":7.6,"publicationDate":"2025-05-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144071682","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}

{"title":"How can we improve data integration to enhance urban air temperature estimations?","authors":"Zitong Wen ,&nbsp;Lu Zhuo ,&nbsp;Meiling Gao ,&nbsp;Dawei Han","doi":"10.1016/j.jag.2025.104599","DOIUrl":"10.1016/j.jag.2025.104599","url":null,"abstract":"<div><div>High-resolution urban air temperatures are indispensable for analysing excess mortality during heatwaves. As a crucial method for obtaining high-resolution data, multi-source data integration has been widely used in urban temperature estimations. However, current research predominantly focuses solely on integrating official weather station observations, satellite products, and reanalysis datasets. Despite the significant cooling effect of rainfall on air temperatures, no studies have explored the contribution of rainfall-related variables to high-resolution air temperature estimations. Additionally, due to the scarcity of official weather stations, quantifying the impact of station density remains an underexplored research direction. To tackle these challenges, we innovatively integrated satellite products, reanalysis datasets, and weather radar data with air temperature observations from crowdsourced weather stations. Using genetic programming, we developed statistical downscaling models to estimate high spatiotemporal resolution (1-km, hourly) air temperatures in London during the summers of 2019 and 2022. The models achieved <em>RMSEs</em> of 1.694 °C (2019) and 1.785 °C (2022), <em>R-squared</em> values of 0.867 and 0.862, and <em>MAEs</em> of 1.276 °C and 1.278 °C, respectively. Notably, the accuracy of the models was found to improve with increased weather station density, particularly when the density was below 0.5 stations per 100 km². Moreover, high-resolution rainfall observations significantly impacted the accuracy of air temperature estimations, second only to elevation, highlighting the potential of integrating radar data. These findings can provide valuable insights for scholars aiming to improve data integration for enhancing urban air temperature estimations.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"140 ","pages":"Article 104599"},"PeriodicalIF":7.6,"publicationDate":"2025-05-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144071683","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}

{"title":"Advancing sustainable agricultural: A novel framework for mapping annual 30-m resolution national-level irrigated croplands","authors":"Enyu Du ,&nbsp;Fang Chen ,&nbsp;Huicong Jia ,&nbsp;Jinwei Dong ,&nbsp;Lei Wang ,&nbsp;Yu Chen","doi":"10.1016/j.jag.2025.104589","DOIUrl":"10.1016/j.jag.2025.104589","url":null,"abstract":"<div><div>With regard to climate change and population growth, irrigated croplands need to be accurately delineated for sustainable water resource management. Owing to the lack of extensive training samples and the limitations of coarse spatiotemporal resolution data in complex agricultural regions, China’s irrigated croplands are difficult to map with a unified spatiotemporal framework. This study presents an innovative method for mapping irrigated and rainfed croplands in mainland China with a local adaptive random forest classifier on the Google Earth Engine platform. Based on the dynamic threshold extraction of multiple peak vegetation index values and a rigorous multi-dataset integration strategy, the annual sample sets of irrigated and rainfed croplands are generated automatically. After constructing 147 multi-feature variables sensitive to irrigation activities, China’s annual irrigated croplands dataset (CAICD) is developed, with 30-m spatial resolution for the 1990–2022 period. The results show the following:(1) CAICD has higher accuracy and a more realistic spatial distribution of irrigated croplands compared with existing datasets, with an average overall accuracy of 0.80. (2) The most sensitive classification features for irrigation signals are spectral indices and original bands, with regional differences influenced by climate characteristics (precipitation and evapotranspiration) and terrain features. (3) Over the past three decades, China’s irrigated croplands have expanded overall and Xinjiang has exhibited the most significant increase and the highest growth rate of irrigated area in mainland China, with an annual expansion of 103 thousand hectares. The results exhibit significant implications for the balance between food security and water resource security, providing valuable insights and contributions for future global monitoring of irrigated croplands.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"140 ","pages":"Article 104589"},"PeriodicalIF":7.6,"publicationDate":"2025-05-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144069700","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}

{"title":"A new method for rapid construction of multi-band vegetation index","authors":"Chenbo Yang ,&nbsp;Juan Bai ,&nbsp;Hui Sun ,&nbsp;Rutian Bi ,&nbsp;Lifang Song ,&nbsp;Chao Wang ,&nbsp;Yu Zhao ,&nbsp;Wude Yang ,&nbsp;Lujie Xiao ,&nbsp;Meijun Zhang ,&nbsp;Xiaoyan Song ,&nbsp;Meichen Feng","doi":"10.1016/j.jag.2025.104601","DOIUrl":"10.1016/j.jag.2025.104601","url":null,"abstract":"<div><div>The aboveground dry biomass (AGDB), leaf area index (LAI), chlorophyll density (CHD), and plant water content (PWC) are important growth physiological parameters that can reflect the growth status of winter wheat. Monitoring the growth physiological parameters of winter wheat by constructing vegetation index is a common method that can quickly obtain the growth physiological parameters of winter wheat. In order to further improve the accuracy of using vegetation index method to monitor the growth physiological parameters of winter wheat, this study proposed a new method based on the genetic algorithm that can quickly construct multi-band vegetation index. The main results were as follows: Compared with the traditional random bands combination to construct the vegetation index, the results of using the rapid method proposed in this study to construct the vegetation index were reliable, and can greatly reduce the time required to construct the vegetation index, which only took about 17.7044 s on average. In the four arithmetic operations used, the priority should be given to Subtraction(Sub.) and Division(Div.), combined with Addition(Add.) and Multiplication(Mul.) operations to construct multi-band vegetation index can achieve good monitoring results. In this study, AGDB, LAI, CHD, and PWC reached the highest accuracy when using ’Sub.-Add.-Sub.-Sub., ’Div.-Div.-Mul.-Mul.’, ’Sub.-Add.-Sub.-Add.’, and ’Div.-Sub.-Div.-Add.’ to construct the five-band vegetation index, respectively, with validation R² (R²_v) of 0.6104, 0.6849, 0.7019, and 0.8960, and validation RMSE (RMSE_v) of 0.4508, 2.1025, 1.1473, and 3.4683. Using ’Sub.-Add.-Sub.-Sub.’ to construct five-band vegetation index could simultaneously monitor four growth physiological parameters. The R²_v of AGDB, LAI, CHD, and PWC monitored by it was 0.6104, 0.6326, 0.6791, and 0.8425 respectively, and the RMSE_v was 0.4508, 2.2702, 1.1903, and 4.2692, respectively. In a word, using the rapid method proposed in this study to construct vegetation index can not only greatly reduce the time required to construct vegetation index, but also has high reliability and reliable operation results.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"140 ","pages":"Article 104601"},"PeriodicalIF":7.6,"publicationDate":"2025-05-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143947108","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}

{"title":"From spark to suppression: An overview of wildfire monitoring, progression prediction, and extinguishing techniques","authors":"Guangqing Zhai ,&nbsp;Longhui Dou ,&nbsp;Yifan Gu ,&nbsp;Hongyi Zhou ,&nbsp;Lele Feng ,&nbsp;Liangliang Jiang ,&nbsp;Jie Dong ,&nbsp;Jiaxuan Sun ,&nbsp;Haidong Li","doi":"10.1016/j.jag.2025.104600","DOIUrl":"10.1016/j.jag.2025.104600","url":null,"abstract":"<div><div>Wildfires, a natural phenomenon predating human civilization, present severe threats to ecosystems, socio-economic factors, and human health. Due to climate change and human influence, the frequency and intensity of global wildfires are on the rise, which emit gigantic amounts of emissions into the atmosphere and compound the world’s efforts to tackle global warming. This study introduces and summarizes the response to wildfires, including fire monitoring, development prediction, and firefighting technology. Satellites, watchtowers, drones, and wireless sensor networks provide comprehensive forest fire monitoring data to fire departments. Artificial intelligence algorithms enhance data analysis and processing efficiency. Real-time wildfire risk prediction strategically guides fire force deployment, optimizing limited resources. The use of unmanned equipment in frontline firefighting enhances efficiency while minimizing risks to firefighters. However, it is revealed that effective rapid control plans for remote, isolated extreme fires remain lacking. This article aims to summarize the latest available technologies and strategies for responding to sudden wildfires, and aid relevant departments and personnel in devising emergency plans for the rapid detection and suppression of wildfires.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"140 ","pages":"Article 104600"},"PeriodicalIF":7.6,"publicationDate":"2025-05-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144069699","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}

{"title":"HMS-PAU-IN: A heterogeneous multi-scale spatiotemporal interaction network model for population analysis units","authors":"Xiaorui Yang ,&nbsp;Rui Li ,&nbsp;Jing Xia ,&nbsp;Junhao Wang ,&nbsp;Hongyan Li ,&nbsp;Nixiao Zou","doi":"10.1016/j.jag.2025.104565","DOIUrl":"10.1016/j.jag.2025.104565","url":null,"abstract":"<div><div>Population analysis units (PAUs), as fundamental spatial units accommodating population-related activities, hold significance in constructing spatiotemporal interaction networks to understand intra-unit population distribution and activity patterns as well as inter-unit interactions. However, existing networks are constrained by fixed scales and the absence of temporal dynamics, with insufficient consideration of multi-scale features, thereby limiting their semantic representation and dynamic analysis capabilities. Thus, we proposed a heterogeneous multi-scale PAU interaction network (HMS-PAU-IN) model that integrates spatial, temporal, and semantic representations, enabling HMS-PAU-IN modeling and semantic analysis based on spatiotemporal knowledge graph. In the spatial dimension, spatiotemporal interactions are classified into explicit interactions driven by population flows and potential interactions shaped by spatial relationships. In the temporal dimension, the changes of PAUs are captured through the evolutionary relationships of nodes between different time windows. To validate the model, we developed a population prediction model that integrates the multi-scale features of PAUs and introduced Leiden-IES-PMS, a community detection method based on the Leiden algorithm, which integrates internal and external environmental semantics and adopts a proximity merging strategy. Experimental results demonstrate that the proposed model and method effectively characterize spatiotemporal interactions among multi-scale PAUs, enhancing the accuracy of population distribution prediction (R² = 0.77) at the community scale, and improving the interpretability of temporal community analysis at the building scale. This study develops a multi-scale spatiotemporal framework for analyzing population distribution, activity patterns, and community evolution within PAUs, providing actionable insights for urban planning, resource optimization, and sustainable management.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"140 ","pages":"Article 104565"},"PeriodicalIF":7.6,"publicationDate":"2025-05-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143942825","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}

{"title":"Soil organic carbon estimation using spaceborne hyperspectral composites on a large scale","authors":"Xiangyu Zhao ,&nbsp;Zhitong Xiong ,&nbsp;Paul Karlshöfer ,&nbsp;Nikolaos Tziolas ,&nbsp;Martin Wiesmeier ,&nbsp;Uta Heiden ,&nbsp;Xiao Xiang Zhu","doi":"10.1016/j.jag.2025.104504","DOIUrl":"10.1016/j.jag.2025.104504","url":null,"abstract":"&lt;div&gt;&lt;div&gt;Soil Organic Carbon (SOC) is a key property for soil health. Spectral reflectance such as multispectral and hyperspectral data could provide efficient and cost-effective retrieval of SOC content. However, constrained by the availability of hyperspectral satellite data, current works mostly use a small number of spaceborne hyperspectral imagery for SOC retrieval on a small scale. In this work, the first large-scale hyperspectral imaging reflectance composites were built, and they were used for SOC estimation. Specifically, DESIS satellite images were used to predict SOC over the whole state of Bavaria in Germany (&lt;span&gt;&lt;math&gt;&lt;mo&gt;∼&lt;/mo&gt;&lt;/math&gt;&lt;/span&gt; 70,000 km&lt;span&gt;&lt;math&gt;&lt;msup&gt;&lt;mrow&gt;&lt;/mrow&gt;&lt;mrow&gt;&lt;mn&gt;2&lt;/mn&gt;&lt;/mrow&gt;&lt;/msup&gt;&lt;/math&gt;&lt;/span&gt;). We prepare 850 hyperspectral images from the DESIS satellite and build temporal composites from them. For the soil data, data was gathered from LfU(Bavarian State Office for the Environment), LfL(Bavarian State Research Center for Agriculture) and LUCAS 2018 (Land Use and Coverage Area Frame Survey). 828 soil samples were selected after data filtering. For this regression task, different machine learning and deep learning methods were implemented and explored. Moreover, a spectral attention mechanism was added to the model. Besides hyperspectral input, the digital elevation model (DEM) was also included as an auxiliary input as the measured spectrum has inter-variability dependent on the elevation and the generated topographical features are also relevant with SOC distribution. Based on the regression results evaluated by &lt;span&gt;&lt;math&gt;&lt;mrow&gt;&lt;mi&gt;R&lt;/mi&gt;&lt;mi&gt;M&lt;/mi&gt;&lt;mi&gt;S&lt;/mi&gt;&lt;mi&gt;E&lt;/mi&gt;&lt;/mrow&gt;&lt;/math&gt;&lt;/span&gt;, &lt;span&gt;&lt;math&gt;&lt;msup&gt;&lt;mrow&gt;&lt;mi&gt;R&lt;/mi&gt;&lt;/mrow&gt;&lt;mrow&gt;&lt;mn&gt;2&lt;/mn&gt;&lt;/mrow&gt;&lt;/msup&gt;&lt;/math&gt;&lt;/span&gt;, and &lt;span&gt;&lt;math&gt;&lt;mrow&gt;&lt;mi&gt;R&lt;/mi&gt;&lt;mi&gt;P&lt;/mi&gt;&lt;mi&gt;I&lt;/mi&gt;&lt;mi&gt;Q&lt;/mi&gt;&lt;/mrow&gt;&lt;/math&gt;&lt;/span&gt;, the deep learning models showed much better performance than machine learning methods. Especially when only using hyperspectral data as input, the best result was achieved with &lt;span&gt;&lt;math&gt;&lt;mrow&gt;&lt;mi&gt;R&lt;/mi&gt;&lt;mi&gt;M&lt;/mi&gt;&lt;mi&gt;S&lt;/mi&gt;&lt;mi&gt;E&lt;/mi&gt;&lt;/mrow&gt;&lt;/math&gt;&lt;/span&gt; 1.947%, &lt;span&gt;&lt;math&gt;&lt;msup&gt;&lt;mrow&gt;&lt;mi&gt;R&lt;/mi&gt;&lt;/mrow&gt;&lt;mrow&gt;&lt;mn&gt;2&lt;/mn&gt;&lt;/mrow&gt;&lt;/msup&gt;&lt;/math&gt;&lt;/span&gt; 0.626, and &lt;span&gt;&lt;math&gt;&lt;mrow&gt;&lt;mi&gt;R&lt;/mi&gt;&lt;mi&gt;P&lt;/mi&gt;&lt;mi&gt;I&lt;/mi&gt;&lt;mi&gt;Q&lt;/mi&gt;&lt;/mrow&gt;&lt;/math&gt;&lt;/span&gt; 1.710 on the test set. After incorporating topographical features, the fused model achieved further improved performance with &lt;span&gt;&lt;math&gt;&lt;mrow&gt;&lt;mi&gt;R&lt;/mi&gt;&lt;mi&gt;M&lt;/mi&gt;&lt;mi&gt;S&lt;/mi&gt;&lt;mi&gt;E&lt;/mi&gt;&lt;/mrow&gt;&lt;/math&gt;&lt;/span&gt; 1.752% and &lt;span&gt;&lt;math&gt;&lt;msup&gt;&lt;mrow&gt;&lt;mi&gt;R&lt;/mi&gt;&lt;/mrow&gt;&lt;mrow&gt;&lt;mn&gt;2&lt;/mn&gt;&lt;/mrow&gt;&lt;/msup&gt;&lt;/math&gt;&lt;/span&gt; 0.695 and &lt;span&gt;&lt;math&gt;&lt;mrow&gt;&lt;mi&gt;R&lt;/mi&gt;&lt;mi&gt;P&lt;/mi&gt;&lt;mi&gt;I&lt;/mi&gt;&lt;mi&gt;Q&lt;/mi&gt;&lt;/mrow&gt;&lt;/math&gt;&lt;/span&gt; 1.919. From the interpretability analysis for model performance, it was found out that the bands in the range of 530 nm–570 nm, 770 nm–790 nm, and 840 nm - 870 nm are the most relevant bands for SOC estimation. In the end, several SOC maps were generated and analyzed together with soil types. The SOC maps indicate that wat","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"140 ","pages":"Article 104504"},"PeriodicalIF":7.6,"publicationDate":"2025-05-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143942826","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}

{"title":"3DLST: 3D Learnable Supertoken Transformer for LiDAR point cloud scene segmentation","authors":"Dening Lu ,&nbsp;Linlin Xu ,&nbsp;Jun Zhou ,&nbsp;Kyle (Yilin) Gao ,&nbsp;Jonathan Li","doi":"10.1016/j.jag.2025.104572","DOIUrl":"10.1016/j.jag.2025.104572","url":null,"abstract":"<div><div>3D Transformers have achieved great success in point cloud understanding and representation. However, there is still considerable scope for further development in effective and efficient Transformers for large-scale LiDAR point cloud scene segmentation. This paper proposes a novel 3D Transformer framework, named <strong>3D L</strong>earnable <strong>S</strong>upertoken <strong>T</strong>ransformer (<strong>3DLST</strong>). The key contributions are summarized as follows. Firstly, we introduce the first Dynamic Supertoken Optimization (DSO) block for efficient token clustering and aggregating, where the learnable supertoken definition avoids the time-consuming pre-processing of traditional superpoint generation. Since the learnable supertokens can be dynamically optimized by multi-level deep features during network learning, they are tailored to the semantic homogeneity-aware token clustering. Secondly, an efficient Cross-Attention-guided Upsampling (CAU) block is proposed for token reconstruction from optimized supertokens. Thirdly, the 3DLST is equipped with a novel W-net architecture instead of the common U-net design, which is more suitable for Transformer-based feature learning. The SOTA performance on challenging LiDAR datasets (airborne MultiSpectral LiDAR (MS-LiDAR) (89.3% of the average <span><math><msub><mrow><mi>F</mi></mrow><mrow><mn>1</mn></mrow></msub></math></span> score) and DALES (80.2% of mIoU)) demonstrate the superiority of 3DLST. Furthermore, 3DLST also achieves satisfactory results in terms of algorithm efficiency, which is up to 5<span><math><mo>×</mo></math></span> faster than previous best-performing methods.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"140 ","pages":"Article 104572"},"PeriodicalIF":7.6,"publicationDate":"2025-05-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143947106","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}

{"title":"Exploring the optimal method for quantifying the contribution of driving factors of urban floods","authors":"Xi Huang ,&nbsp;Linhan Yang ,&nbsp;GuiCheng Yang ,&nbsp;Jiufeng Li ,&nbsp;Luo Liu ,&nbsp;Yilun Liu ,&nbsp;Xianzhe Tang","doi":"10.1016/j.jag.2025.104597","DOIUrl":"10.1016/j.jag.2025.104597","url":null,"abstract":"<div><div>Amid climate change and urban expansion, urban floods (UFs) have become a significant concern. Effective UF management requires two key elements: <strong>(1)</strong> assessing susceptibility to identify flood-prone hotspots and <strong>(2)</strong> analyzing factor contributions to pinpoint primary drivers. While research on urban flood susceptibility (UFS) is well-established, studies on factor contributions usually limited to the simple application without fully evaluating their alignment with UF characteristics. This study addresses this gap by examining Shenzhen, a rapidly urbanizing city, and applying various regression methods (linear, multiple machine learning (ML), and Shapley Additive exPlanations (SHAP)) to evaluate flood-driving factors. Through accuracy assessments and environmental considerations, we identify the most suitable methods for this analysis. Our findings include: <strong>(1)</strong> Random Forest (RF)-based UFS assessment reveals significant spatial heterogeneity, with higher UFS in central-western areas and lower risks in the southeast; <strong>(2)</strong> the most contributing factors in Shenzhen exhibit non-linear and non-parametric effects on UF, supporting the use of RF; <strong>(3)</strong> we found substantial spatial heterogeneity in factor contributions, necessitating SHAP analysis to capture local variations and provide tailored management insights. Thus, this study integrates methodological rigor with urban environmental context, enhancing factor contribution analysis for UF and supporting effective management strategies.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"140 ","pages":"Article 104597"},"PeriodicalIF":7.6,"publicationDate":"2025-05-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143942897","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}