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

{"title":"Quantifying key indicators of essential biodiversity variables for mangrove species in response to hydro-meteorological factors","authors":"Hang Yao ,&nbsp;Bolin Fu ,&nbsp;Weiwei Sun ,&nbsp;Yuyu Zhou ,&nbsp;Yeqiao Wang ,&nbsp;Weiguo Jiang ,&nbsp;Hongchang He ,&nbsp;Zhili Chen ,&nbsp;Yiji Song","doi":"10.1016/j.jag.2025.104535","DOIUrl":"10.1016/j.jag.2025.104535","url":null,"abstract":"<div><div>Mangroves are critical for climate mitigation and biodiversity conservation, yet their spatiotemporal dynamics and physiological responses to hydrometeorological drivers remain poorly understood. This study extracted three essential biodiversity variables (area distribution, phenology, and physiological traits) and further revealed their dependencies on hydrometeorological conditions. We developed a continuous time-series monitoring method (CTSM) to enhance the Detect-Monitor-Predict detection framework for accurately tracking mangrove spatial succession in the Beibu Gulf from 2000 to 2021. We combined Continuous Change Detection and Classification with Harmonic Analysis of Time Series (HANTS) methods to capture the seasonal changes of physiological traits of dominant mangrove species. This study utilized HANTS-PLSR (partial least squares regression) response models and structural equation models to explore the seasonal responses of physiological trait to hydro-meteorological factors. The results indicated that (1) the improved detect component delineated fine-scale expansion patterns of mangroves, with area-hydrometeorology coupling evolving from uncoordinated to highly coordination during 2000–2021. (2) The start, peak and end of the growing season for mangroves are in March-April, June-September and January-February of the following year, respectively. The mangroves in different regions exhibit relatively delayed growth periods. (3) <em>Aegiceras corniculatum</em> exhibited bimodal phenological trajectories, contrasting with unimodal patterns in three co-occurring species. (4) The physiological traits displayed a positive correlation with water/air temperature and sunshine duration. The phenological changes of four mangrove species are driven by the interaction between hydrological and meteorological variables, with meteorological factors dominating (path coefficient &gt; 0.50, <em>p</em> &lt; 0.001). The findings provide insights into mangrove conservation and biodiversity monitoring.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"139 ","pages":"Article 104535"},"PeriodicalIF":7.6,"publicationDate":"2025-04-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143842999","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":"Leveraging moisture elimination and hybrid deep learning models for soil organic carbon mapping with multi-modal remote sensing data","authors":"Yilin Bao ,&nbsp;Xiangtian Meng ,&nbsp;Weimin Ruan ,&nbsp;Huanjun Liu ,&nbsp;Mingchang Wang ,&nbsp;Abdul Mounem Mouazen","doi":"10.1016/j.jag.2025.104513","DOIUrl":"10.1016/j.jag.2025.104513","url":null,"abstract":"<div><div>Precision management of soil organic carbon (SOC) is crucial for regulating the global carbon cycle and ensuring food security. Currently, SOC prediction remains challenging due to unresolved moisture disturbances, underutilized multimodal remote sensing data, and uncertain model transferability. To address these challenges, a new paradigm integrating moisture elimination with advanced hybrid deep learning has been developed. Firstly, this research employs optimal cluster analysis based on soil moisture spatial characteristics, followed by the application of the Kubelka-Munk radiative transfer model to construct a moisture elimination strategy (MES). Next, a hybrid deep learning model, Multimodal Transformer Mechanism-Convolutional Neural Network-Convolutional Long Short-Term Memory (MT-CNN-ConvLSTM, MCCL), is constructed to enhance predictive accuracy and generalizability. The MCCL model was compared to other machine learning and deep learning models, including LSTM, Random Forest (RF), CNN, Artificial Neural Network (ANN), Support Vector Machine (SVM) and partial least squares regression (PLSR). Results indicate that (1) the proposed paradigm achieves optimal SOC content prediction accuracy in humid regions, with a root mean square error (RMSE) of 3.58 g kg⁻¹, a coefficient of determination (R²) of 0.76, a ratio of performance to interquartile distance (RPIQ) of 2.26, and a mean absolute error (MAE) of 4.73 g kg⁻¹. The model shows better performance in semi-humid regions, yielding an RMSE of 3.12 g kg⁻¹, R² of 0.77, RPIQ of 2.27, and MAE of 4.71 g kg⁻¹, indicating significant spatial transferability. (2) Under MES, multiple models showed improved R² using PLSR as the baseline: e.g., MCCL (41.4 %), LSTM (28.3 %), RF (17.2 %), CNN (14.1 %), ANN (8.1 %), and SVM (7.1 %). (3) The MES approach reduces RMSE by 1.06 g kg⁻¹ and MAE by 1.58 g kg⁻¹, while increasing R² by 18.75 %, and RPIQ by 0.82. Using the KM radiative transfer model without cluster partitioning decreases RMSE and MAE by 0.58 g kg⁻¹ and 0.23 g kg⁻¹, while increasing R² and RPIQ by 7.1 % and 0.3, respectively. Specifying the soil moisture gradient in the spectral correction process is crucial. The novel MES-MCCL paradigm proposed in this study is robust and provides promising insights into soil moisture masking’s spectral characterization and the potential of multimodal remote sensing for SOC monitoring.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"139 ","pages":"Article 104513"},"PeriodicalIF":7.6,"publicationDate":"2025-04-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143843040","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":"Extraction of built-up areas using Sentinel-1 and Sentinel-2 data with automated training data sampling and label noise robust cross-fusion neural networks","authors":"Yu Li,&nbsp;Patrick Matgen,&nbsp;Marco Chini","doi":"10.1016/j.jag.2025.104524","DOIUrl":"10.1016/j.jag.2025.104524","url":null,"abstract":"<div><div>Up-to-date mapping of built-up areas is of paramount importance for urban planning, environmental monitoring, and disaster management. In recent years, there has been a growing interest in employing supervised machine learning and deep learning methods to map built-up areas using satellite SAR and optical data. However, the laborious and expensive task of gathering and maintaining a vast array of diverse training data poses a challenge to the widespread adoption of these methods for large-scale built-up area mapping. This paper presents a two-step framework enabling an automated extraction of built-up areas using Sentinel-1 and Sentinel-2 data. Initially, training data for built-up and non-built-up classes are automatically sampled and labeled from Sentinel-1 and Sentinel-2 data for a given area of interest. Subsequently, a cross-fusion neural network is trained using the samples from the first step to produce a built-up map for the entire study area. To enhance the network’s resilience to label noise, a contextual virtual adversarial training (CVAT) regularization is introduced within the mean-teacher architecture. Our proposed framework was tested on 48 different study areas across the world. Both quantitative and qualitative evaluations demonstrate its robustness and effectiveness for large-scale built-up area extraction. The versatility of our framework in generating accurate and up-to-date built-up information, which is essential for monitoring urban environments and assessing economic losses resulting from natural disasters, is highlighted through comparisons with four state-of-the-art global built-up products: Global Human Settlement Built-up map based on 2018 Sentinel-2 composites (GHS-BUILT-S2), World Settlement Footprint 2019 (WSF 2019), ESA World Cover, and Dynamic World.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"139 ","pages":"Article 104524"},"PeriodicalIF":7.6,"publicationDate":"2025-04-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143826039","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":"Enhancing global aerosol retrieval from satellite data via deep learning with mutual information estimation","authors":"Xiaohu Sun ,&nbsp;Yong Xue ,&nbsp;Lin Sun","doi":"10.1016/j.jag.2025.104534","DOIUrl":"10.1016/j.jag.2025.104534","url":null,"abstract":"<div><div>Satellite-based data can provide continuous aerosol observations but suffer from significant uncertainties across various regions. Transfer learning improves model generalization, yet its application in atmospheric research remains limited. Here, we introduce an innovative framework for retrieving global aerosol optical depth (AOD) which named the <strong>A</strong>erosol domain-<strong>Ada</strong>ptive <strong>N</strong>etwork (AAdaN). The framework utilizes a neural network to estimate mutual information, and aligns spatial covariate shift via a transfer loss term. Then, we assess the retrieval potential in unknown scenarios using independent land cover type, and the proposed model demonstrates satisfactory results. The cross-validation shows strong agreement with in-situ measurements, both in sample-based and site-based evaluations. Specifically, the site-based ten-fold cross-validation of our AOD retrievals indicate that all accuracy metrics are satisfactory, with a Pearson correlation of 0.766 and a Root-Mean-Square Error of 0.118, and that about 76.05 % of the retrievals meet the expected error criteria [±(0.05 + 20 %)]. Additionally, the proposed AAdaN achieves stable, high-accuracy aerosol retrievals across various surface and atmospheric conditions, and can generate spatially continuous AOD distributions. This study significantly improves spatial generalization and offers valuable insights for future model development.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"139 ","pages":"Article 104534"},"PeriodicalIF":7.6,"publicationDate":"2025-04-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143829875","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":"Snow depth inversion and mapping at 500 m resolution from 1980 to 2020 in Northeast China using radiative transfer model and machine learning","authors":"Yanlin Wei ,&nbsp;Xiaofeng Li ,&nbsp;Lingjia Gu ,&nbsp;Zhaojun Zheng ,&nbsp;Xingming Zheng ,&nbsp;Tao Jiang","doi":"10.1016/j.jag.2025.104533","DOIUrl":"10.1016/j.jag.2025.104533","url":null,"abstract":"<div><div>Accurate snow cover parameter assessment and mapping at a fine resolution can have profound implications for our understanding of the planet’s water balance and energy dynamics. Passive microwave (PMW) remote sensing is among the most effective methods for snow depth (SD) retrieval. However, significant uncertainties persist in SD retrieval and mapping due to snow characteristic variations, forest canopy interference, and low spatial resolution of PMW. To overcome these limitations, a novel method considering multiple influencing factors was proposed by integration a radiation transfer model with a machine learning model for SD retrieval, and a 500 m resolution SD dataset (NCSD) was generated for 1980 − 2020 in Northeast China by combining downscaling model. The validation against independent ground observations revealed that the <em>MAE</em>, <em>RMSE</em>, and <em>R</em> values for NCSD were 4.39 cm, 6.65 cm and 0.77, respectively. Compared to existing SD products, NCSD data effectively avoid mixed pixel issues, improved SD retrieval performance, and reveal more refined snow cover spatiotemporal patterns. Additionally, the NCSD results indicated that the annual average SD in Northeast China exhibited an increasing trend from 1980 to 2020 (0.26 cm/10a, <em>p</em> &gt; 0.05). However, a notable inflection point occurred in 2000, and a subsequent decreasing trend occurred from 2000 to 2020 (0.49 cm/10a, <em>p</em> &gt; 0.05). Overall, the creation of NCSD effectively filled the gap related to high-resolution SD data, and facilitated the development of hydrological studies and climate change at the basin scale.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"139 ","pages":"Article 104533"},"PeriodicalIF":7.6,"publicationDate":"2025-04-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143823577","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 10-meter global terrace mapping using sentinel-2 imagery and topographic features with deep learning methods and cloud computing platform support","authors":"Yifan Li ,&nbsp;Fuyou Tian ,&nbsp;Miao Zhang ,&nbsp;Hongwei Zeng ,&nbsp;Shukri Ahmed ,&nbsp;Xinli Qin ,&nbsp;Yanxu Liu ,&nbsp;Lizhe Wang ,&nbsp;Runyu Fan ,&nbsp;Bingfang Wu","doi":"10.1016/j.jag.2025.104528","DOIUrl":"10.1016/j.jag.2025.104528","url":null,"abstract":"<div><div>Terrace agriculture plays a vital role in mountainous regions by preventing soil erosion, optimizing land use, and supporting local ecosystems. However, research on the global distribution of terraces is limited due to the lack of unified automatic identification model. Despite the rapid advancements in deep-learning architectures in recent years, their performance in extracting terrace maps still needs investigation. To address this limitation, this study compares the performance of eight state-of-the-art deep learning models, including UNet, HRNet, DeepLabv3+, TransUNet, Segmenter, PVT v2, Swin-Unet, and PerSAM. Sentinel-2 imagery was selected for its spectral properties, while Digital Elevation Model (DEM) imagery was chosen for detailed topographic information. UNet outperformed others in terrace identification, achieving an overall accuracy of 92.8 % and a mean Intersection over Union (MIoU) of 75.9 %. The entire data processing workflow, using Google Earth Engine for data acquisition, Google Drive for storage, Google Earth Pro for computational capabilities, and the T4 GPU in cloud computing resources, requires approximately 625 h. As a result, the Global Terrace Map (GTM) was generated at 10-meter resolution for 2022. The total terrace area was estimated at 853,161 km2, accounting for about 5.1 % of global cropland. The countries with the most extensive terraced areas, as identified, are China (298,908 km², 18 % of total cropland), Ethiopia (127,266 km², 47 %), Kenya (36,385 km², 37 %), India (34,485 km², 2 %), and Democratic Republic of the Congo (31,422 km², 21 %). This pioneering global terrace map is anticipated to bridge a significant data gap in the field of resilient agriculture. It will offer invaluable insights into the spatial distribution and attributes of terraced farming systems, along with their roles in enhancing food security and promoting environmental sustainability.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"139 ","pages":"Article 104528"},"PeriodicalIF":7.6,"publicationDate":"2025-04-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143823267","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":"Local pathways of association","authors":"Jiao Hu ,&nbsp;Rui Qu ,&nbsp;Yongze Song ,&nbsp;Peng Wu","doi":"10.1016/j.jag.2025.104531","DOIUrl":"10.1016/j.jag.2025.104531","url":null,"abstract":"<div><div>Spatial association reveals the interconnected nature of geographical phenomena, describing the interactions and influences of environmental variables across geographic space. Path analysis can explore complex causal relationships between variables by analyzing path coefficients. However, in large-scale studies, path analysis methods are often affected by local effects, which can influence the accuracy and reliability of the results. This study develops a local pathway association (LPA) model to analyze local effects of pathways among variables that integrates path analysis and local pathway coefficient estimations. The LPA model was employed to investigate the spatial heterogeneity of spatial associations between factors such as climate, soil, and vegetation on the Tibetan Plateau. Results indicate that the LPA model effectively reveals the spatial variation characteristics of local path coefficients between geographic variables, avoiding the underestimation or overestimation of global path coefficients in traditional path coefficient studies. The developed LPA model provides an effective technical tool for revealing spatial differences in path associations of large-scale spatial studies. The strong data compatibility of the LPA model allows for broad applicability across various disciplines and a deeper understanding of localized interactions and variations in complex geospatial and Earth systems.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"139 ","pages":"Article 104531"},"PeriodicalIF":7.6,"publicationDate":"2025-04-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143823268","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":"GPP estimation by transfer learning with combined solar-induced chlorophyll fluorescence and eddy covariance data","authors":"Yongming Ma ,&nbsp;Xiaobin Guan ,&nbsp;Yuchen Wang ,&nbsp;Yuyu Li ,&nbsp;Dekun Lin ,&nbsp;Huanfeng Shen","doi":"10.1016/j.jag.2025.104503","DOIUrl":"10.1016/j.jag.2025.104503","url":null,"abstract":"<div><div>Gross primary productivity (GPP) plays a crucial role in the carbon exchange between the atmosphere and terrestrial ecosystems. Eddy covariance (EC) method can obtain accurate GPP at the site level, but the sparse distribution limits representativeness. Satellite solar-induced chlorophyll fluorescence (SIF) serves as emerging data of large-scale GPP, yet there are still limitations in its conversion to GPP and spatiotemporal coverage. This study proposes a transfer learning (SIFEC-TL) method to estimate long-term global GPP with high accuracy by combining constraints from SIF and EC data. SIF data are taken as the source domain that provides the spatial information for pre-training, and EC GPP in the target domain provides precise GPP for the machine learning model fine-tuning. To verify the performance of SIFEC-TL, the results are compared with those from machine learning models that use only SIF or EC GPP alone (SIFML and ECML). The results indicate that the SIFEC-TL model demonstrates stronger spatial scalability compared to the SIFML and ECML models, with R² increasing by 0.132 and 0.036. The SIFEC-TL more effectively captures inter-annual GPP dynamics with underestimation/overestimation over high/low values in the SIFML and ECML models being well corrected. Furthermore, three different SIF-based GPP are also used as source domains, and the results showed that they only affect pre-training but the final accuracy after fine-tuning remains similar, which indicates SIFEC-TL can obtain stable GPP estimation accuracy regardless of the spatiotemporal coverage of SIF data and its conversion to GPP.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"139 ","pages":"Article 104503"},"PeriodicalIF":7.6,"publicationDate":"2025-04-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143823269","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":"More intelligent knowledge graph: A large language model-driven method for knowledge representation in geospatial digital twins","authors":"Jinbin Zhang ,&nbsp;Jun Zhu ,&nbsp;Zhihao Guo ,&nbsp;Jianlin Wu ,&nbsp;Yukun Guo ,&nbsp;Jianbo Lai ,&nbsp;Weilian Li","doi":"10.1016/j.jag.2025.104527","DOIUrl":"10.1016/j.jag.2025.104527","url":null,"abstract":"<div><div>Knowledge graphs (KGs) can describe the nature and relationships of geographic entities and are an essential knowledge base for realizing geospatial digital twins (GDTs). However, existing KGs make it challenging to describe dynamic geographic entities under geographic spatiotemporal evolution accurately. Furthermore, they are constrained by the professional backgrounds of their users, which hinders updates and communication. Therefore, the research constructed an “event-object-state” three-domain associated GDT-oriented KG, proposed a large language model (LLM) −driven KG dynamic update algorithm, and established a KG intelligent Q&amp;A method integrating LLM. We developed a prototype system and selected an earthquake disaster as a typical geographic event for experimental analysis. The results showed that the proposed method can reflect the space, time, state, evolution process, and interrelationships of geographic entities in a more comprehensive way, support users to build, update, and query KGs using natural language, with an updating efficiency of less than 1 min, and an updating quality comparable to that of manual updating by experts. Compared with the traditional KGs, our method can represent virtual geographic entities and has significant advantages in intelligence and automation, which effectively breaks down professional barriers and supports the construction of GDTs with the need for rapid updating of knowledge.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"139 ","pages":"Article 104527"},"PeriodicalIF":7.6,"publicationDate":"2025-04-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143823270","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":"SASTGCN: Semantic-Augmented Spatio-temporal graph convolutional network for subway flow prediction","authors":"Shiyuan Jin ,&nbsp;Changfeng Jing ,&nbsp;Sheng Yao ,&nbsp;Yushan Zhang ,&nbsp;Pu Zhao ,&nbsp;Jinlong Zhang","doi":"10.1016/j.jag.2025.104530","DOIUrl":"10.1016/j.jag.2025.104530","url":null,"abstract":"<div><div>Deep learning based subway passenger flow prediction was widely employed to promote prediction accuracy, which is crucial for subway management and commercial infrastructure planning. However, the existing work ignored the semantic similarity inherent in the subway stations function, which can extract passengers and enhance prediction accuracy. In this work, a Semantic-Augmented Spatio-temporal Graph Convolutional Network (SASTGCN) model was proposed, which considered semantic similarity, spatiotemporal correlations and spatial heterogeneity to realize the passenger inflow and outflow prediction. The station function was derived from travel characteristics of passengers by data-driven method. The spatiotemporal block including Topology Adaptive Graph Convolutional Network (TAGCN) and ConvNeXt, constructed adaptive spatial topology, depthwise separable convolution and expanded receptive fields to capture spatiotemporal correlations and spatial heterogeneity. The SASTGCN model was validated with the card swiping data in Shanghai, the prediction ability and error analysis results demonstrated the performance outperform nine baseline methods, and the accuracy was improved by approximately 21%. The proposed model can provide inspiration for the follow-up research of passenger flow prediction, traffic pattern recognition and dynamic scheduling.</div></div>","PeriodicalId":73423,"journal":{"name":"International journal of applied earth observation and geoinformation : ITC journal","volume":"139 ","pages":"Article 104530"},"PeriodicalIF":7.6,"publicationDate":"2025-04-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143816228","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}