Table of Content

23 June 2025, Volume 54 Issue 5

Previous Issue   

Digital Twin

Smart city logical framework and digital-twin platform technical requirements

Renzhong GUO, Biao HE, Zhigang ZHAO, Xiaoming LI, Xi KUAI, Haojia LIN, Yebin CHEN, Ding MA

2025, 54(5):  777-784.  doi:10.11947/j.AGCS.2025.20240001

Asbtract ( )   HTML ( )   PDF (2439KB) ( )  

Figures and Tables | References | Related Articles | Metrics

Cities are open, complex, and dynamically changing “nature-society” mega-systems, where various functional and structural subsystems continuously interact and influence one another, forming the inherent logic of urban operations. Smart cities present viable solutions to address the broad systemic issues within urban areas. However, the traditional vertical coupling architecture, despite its self-sufficiency, has failed to align effectively with urban operational logic, resulting in numerous “data islands” and “information silos” that impede the process of high-quality urban development. This article adopts an urban system perspective, grounded in the developmental needs of smart cities in the information and communication technology (ICT) era. It proposes a lateral coupling system architecture that facilitates joint construction and sharing for smart city development from theoretical, technological, and methodological standpoints. Furthermore, the study seeks to clarify the concept of the urban system from a philosophical perspective and to identify the key technical requirements for a digital twin platform, thereby providing guidance for the engineering practice and innovative application of smart cities.

Concept and technical system of smart city operating system

Biao HE, Renzhong GUO, Hai XU, Xi KUAI, Haojia LIN, Zhigang ZHAO

2025, 54(5):  785-794.  doi:10.11947/j.AGCS.2025.20240120

Asbtract ( )   HTML ( )   PDF (4080KB) ( )  

Figures and Tables | References | Related Articles | Metrics

The development of smart cities has transitioned from the construction of network communication infrastructure to the management of intelligent urban systems. The core challenge in building smart cities today lies in how to develop urban spatio-temporal data and a unified spatio-temporal computing capability to meet the diverse needs of various applications. This paper introduces and analyzes the concept of smart city operating system, elaborating on its implications through the lenses of urban ternary spatial cognition, logical layering, and system architecture. It also emphasizes the importance of focusing on ubiquitous city sensing and digital modeling, spatio-temporal big data fusion and analysis, three-dimensional urban visualization, city-level simulation, and the integration of a city's computational power network. Additionally, it highlights the necessity of unified scheduling within the smart city operating system's technological framework. This paper aims to offer valuable insights for the development of smart cities in the contemporary era.

Geodesy and Navigation

Prediction method of regional tropospheric wet delay based on Conv-LSTM network

Haopeng FAN, Bojiao ZHANG, Zhongmiao SUN, Jinkai FENG

2025, 54(5):  795-804.  doi:10.11947/j.AGCS.2025.20240108

Asbtract ( )   HTML ( )   PDF (5232KB) ( )  

Figures and Tables | References | Related Articles | Metrics

The tropospheric zenith wet delay (ZWD) is time-varying and varies with geographical locations, which has become one of the main bottlenecks restricting the accuracy or timeliness of various spatial geodetic technologies. In view of this, a prediction method based on convolutional long-short term memory (ConvLSTM) network was exploited, during which a continuation of regional historical ZWD was conducted to enhance the spatial correlation, and an incremental training was adopted to improve the attention of spatio-temporal series to sudden changing signals; finally, taking the central European region as an example, the calculation effects of the sliding window conic extrapolation, the classical ConvLSTM and the method in this paper were compared. The results show that the short-term accuracy of the sliding window conic method is equivalent to that of the classical ConvLSTM; yet, when the prediction span increases, the accuracy of the former decreases sharply, while the latter is almost unaffected. After using the incremental improvement method, the accuracy is improved by 60% on the basis of the classical ConvLSTM method; after employing the “extension+increment” method, the systematic error is even further reduced by more than 50%.

The impact of environmental loading on nonlinear variations of 3D coordinate time series of GNSS stations in Sichuan and Yunnan region

Shunqiang HU, Kejie CHEN, Xiaoxing HE, Hai ZHU, Tan WANG

2025, 54(5):  805-818.  doi:10.11947/j.AGCS.2025.20240397

Asbtract ( )   HTML ( )   PDF (19533KB) ( )  

Figures and Tables | References | Related Articles | Metrics

Investigating the nonlinear variations in GNSS coordinate time series can obtain a high-precision GNSS velocity fields, which contribute to establish and maintain a dynamic Earth reference frame. Here, we used the 94 GNSS stations data over the period January 2011 to December 2022 at Sichuan and Yunnan region and calculated the environmental loading deformation associated with hydrological loading, atmospheric loading, and non-tidal ocean loading, which provided by the IMLS and GFZ, respectively. We quantitatively evaluates the nonlinear changes in the GNSS 3D coordinate time series using PRMS that the percentage of RMS change after environmental loading correction. Finally, the ICA model was carried to separate the nonlinear signals in the vertical GNSS coordinate time series. The results show that the environmental loading products provided by IMLS outperforms the GFZ. The hydrological loading and atmospheric loading significantly correct nonlinear changes in vertical GNSS coordinate time series, with an average PRMS values of 8.66% and 6.74%, respectively. In the horizontal north (N) component, the hydrological loading can weaken the nonlinear changes in the GNSS coordinate time series, with an average PRMS values of 2.07%; while the atmospheric loading will increase the nonlinear changes, with an average PRMS values of -0.22%. In the horizontal east (E) component, the hydrological loading increase the nonlinear changes, with an average PRMS values of -0.11%; while the effect of atmospheric loading correction is weak, with an average PRMS values of 0.22%. For the NEU components, the effect of non-tidal ocean loading correction is not obvious in nonlinear changes of GNSS coordinate time series, with an average PRMS values of 0.33%, -0.2%, and 0.63%, respectively. The correlation coefficients between the IC2 signal separated by the ICA method and the IMLS atmospheric loading deformation range from -0.68 to 0.71 (greater than 0.65 accounting for 73% of all GNSS stations), with an average value of 0.64, while the correlation coefficients between the IC4 signal and the hydrological loading deformation obtained by the GRACE/GRACE-FO model range from 0.55 to 0.79, with an average value of 0.69, indicating that the IC2 and IC4 signals may be related to the nonlinear changes caused by atmospheric loading and hydrological loading, respectively. Furthermore, it is necessary to correct nonlinear changes in GNSS coordinate time series using environmental loading products.

Simulation and verification of regional hydrological gravity effect considering undulating terrain: a case study of the head region of Three Gorges Reservoir

Mingtao ZHU, Yi ZHANG, Xian MA, Linsong WANG

2025, 54(5):  819-830.  doi:10.11947/j.AGCS.2025.20240449

Asbtract ( )   HTML ( )   PDF (6758KB) ( )  

Figures and Tables | References | Related Articles | Metrics

The cutoff of the Three Gorges Dam has led to changes in the region's hydrology and even the Yangtze River basin. Most previous studies using river digital elevation and regional hydrological models focused on analyzing the gravity response of simplified water storage models and regional equivalent water heights, without considering the impact of river slopes and surrounding undulating terrain. This study constructed a water storage load model based on the river boundary of the head area of the Three Gorges Reservoir extracted by the Gaofen-1 (GF-1) satellite image. It used the Delaunay method to triangulate the complex water body surface and surrounding terrain. The regional hydrological gravity and gravity gradient effect were simulated using a high-precision polyhedron external gravity field algorithm. The research results indicate that considering the dynamic variation of water level and its impact on the slope provides a more reasonable simulation of the gravitational effects during the reservoir impoundment process. The relative error concerning absolute gravity measurements is 14%, which represents a significant improvement compared to the previous static reservoir models (with relative errors of 50% and 71% for low and high water levels, respectively). Furthermore, the regional hydrological gravity effect simulation suggests that when calculating gravity changes using global or regional hydrological models, the topography surrounding the observation points should be considered. The high-precision reservoir load modeling and forward simulation results for the Three Gorges Reservoir area presented in this paper will provide important support for long-term regional gravity monitoring, comparative analysis, and hydrological gravity correction. The findings also contribute to further exploration of the dynamic processes triggered by reservoir impoundments, such as landslides and seismic activities in the reservoir area.

Marine Survey

Influence of draft depth error of acoustic ray tracing on underwater positioning

Wenzhou SUN, Anmin ZENG, Zhengming QIAO

2025, 54(5):  831-839.  doi:10.11947/j.AGCS.2025.20230117

Asbtract ( )   HTML ( )   PDF (1470KB) ( )  

Figures and Tables | References | Related Articles | Metrics

Acoustic underwater positioning technology is an important means of obtaining three-dimensional absolute coordinates of seabed control points. The accuracy of the initial depth calculation for acoustic ray tracing is affected by the draft depth error of the ship-borne transducer. The draft depth error causes corresponding system errors in slant range calculations, thereby affecting the positioning results of seafloor control points. This paper first studies the relationship between draft depth error and acoustic ranging error through the area difference of sound velocity profiles. It analyzes the variation pattern of acoustic ranging error caused by draft depth error with respect to gradients. Secondly, it analyzes the variation patterns of acoustic ranging error caused by draft depth error with respect to depth and initial incident angle under typical deep-sea sound velocity profile structures, as well as their impact on circular track positioning method. Simulation experiments indicate that draft depth errors can lead to centimeter or even decimeter-level ranging errors. Real measured data at a depth of 3000 m shows that draft depth errors cause decimeter-level deviations in vertical solutions, and the larger the radius of circular track, the greater the deviation in the vertical solution.

Multipath negative outlier removal method for coastal LiDAR point clouds based on mirror structure and intensity feature constraints

Haolong GAO, Shaobo LI, Jianhu ZHAO

2025, 54(5):  840-852.  doi:10.11947/j.AGCS.2025.20240242

Asbtract ( )   HTML ( )   PDF (6193KB) ( )  

Figures and Tables | References | Related Articles | Metrics

The hovercraft-mounted LiDAR scanning system can acquire coastal zone point cloud data. However, the quality of the data is affected by negative outliers caused by the multipath effects, which limits its wide applications. This paper studied the formation mechanism of such outliers and proposed a negative outlier removal method based on mirror structure and intensity feature constraints. First, a mirror surface was obtained using the region growing and quadratic least squares fitting method. The geometric structure and intensity features of the point cloud clusters identified through region growing were then characterized. After that, using the mirror surface, the relationship of the geometric and intensity features of the point cloud clusters within the same mirror space was established. This relationship facilitated the identification and removal of negative outliers exhibiting mirror characteristics. Experimental results showed that the proposed method achieved a precision of 90.83% and a recall of 76.05% in removing negative outliers from coastal tidal flats. The method can effectively eliminate negative outliers while preserving terrain features, thereby enhancing the reliability of point cloud data for high-precision noise reduction in subsequent applications.

Photogrammetry and Remote Sensing

Visual-language joint representation and intelligent interpretation of remote sensing geo-objects: principles, challenges and opportunities

Haifeng LI, Wang GUO, Mengwei WU, Chengli PENG, Qing ZHU, Yu LIU, Chao TAO

2025, 54(5):  853-872.  doi:10.11947/j.AGCS.2025.20240244

Asbtract ( )   HTML ( )   PDF (5893KB) ( )  

Figures and Tables | References | Related Articles | Metrics

Remote sensing imagery intelligent interpretation primarily relies on visual models to establish a mapping between remote sensing images and semantic labels. However, due to the limited categories of available annotations, such models struggle to capture the deep semantics of geo-objects and their interrelations, thereby failing to develop a broader understanding of world knowledge. With the emergence of large language models (LLMs), which possess powerful capabilities in encoding human knowledge expressed through language, this limitation may be effectively addressed. Guiding visual models with LLMs can significantly broaden their capacity for knowledge acquisition and drive a paradigm shift in remote sensing interpretation—from surface-level semantic matching toward deeper world knowledge understanding. Building upon this insight, this paper presents a systematic analysis of geo-object concept representation in remote sensing. By examining both the intension and extension of geo-object concepts, it reveals the limitations of relying solely on visual modalities for representing complex geo-object characteristics. The study then elaborates on the theoretical significance and practical value of integrating visual and language modalities to enhance concept representation. Furthermore, it investigates the inherent challenges of modality alignment under this new paradigm and reviews representative solution strategies. This paper also explores how the paradigm fosters the emergence of novel capabilities in remote sensing interpretation models, analyzes the underlying mechanisms driving these capabilities, and discusses their practical implications. Finally, it summarizes the new opportunities and challenges that arise in intelligent remote sensing interpretation within this conceptual framework.

Multi-sensor optical remote sensing images change detection based on global differential enhancement module and balance penalty loss

Chao WANG, Tianyu CHEN, Tong ZHANG, Tanvir AHMED, Liqiang JI, Tao XIE, Jiajun YANG, Shuai WANG

2025, 54(5):  873-887.  doi:10.11947/j.AGCS.2025.20240300

Asbtract ( )   HTML ( )   PDF (10472KB) ( )  

Figures and Tables | References | Related Articles | Metrics

High-resolution optical imagery, characterized by its rich spatial details and high interpretability, is one of the primary data sources for remote sensing change detection tasks. In terms of practical applications, single-source optical imagery is often constrained by revisit periods and the availability of archived data, which may not fully meet the requirements. In contrast, the integration of multi-source optical remote sensing imagery offers greater flexibility and applicability. Nonetheless, the substantial spatiotemporal heterogeneity among optical images acquired by different sensors leads to significant “pseudo-invariance” and “pseudo-change” phenomenon, posing severe challenges to accurately extracting genuine change information. In order to address these challenges, this paper aims to develop a model with strong discriminative capability for addressing both types of false detection. In this paper, a change detection model for multi-sensor optical remote sensing images is proposed, based on global differential enhancement module (GDEM) and balance penalty loss (BP Loss), named GB-UNet++. The GDEM introduces Transformer to facilitate the interaction of global information across multi-temporal images, thereby enhancing the model's ability to capture cross-pixel/region change information. Additionally, the proposed BP Loss adaptively adjusts the weights to enhance the model's capacity to learn from both types of misclassified samples. Extensive experiments on six datasets demonstrate that the proposed method achieves an overall accuracy (OA) of 99.02% and an F₁ score of 84.86%, significantly outperforming five state-of-the-art methods.

AOSN: alpha optimal structure network for height estimation from a single SAR image in mountain areas

Qingli LUO, Xueyan LI, Guoman HUANG, Honghui CHEN, Minglong XUE, Jian LI

2025, 54(5):  888-898.  doi:10.11947/j.AGCS.2025.20230431

Asbtract ( )   HTML ( )   PDF (10163KB) ( )  

Figures and Tables | References | Related Articles | Metrics

Height estimation from a single synthetic aperture radar (SAR) image is a possible way to estimate height in all day and all weather conditions with real time capabilities. However, it is an ill-posed problem since the same 2D images may be projected from multiple 3D images. The development of deep learning provides a possible solution for it. The problems of the current deep learning methods are lack of detail feature information and the accuracy of the estimated height is not high enough. In order to address the above issue, this paper proposes alpha optimal structure network (AOSN), utilizing the characteristics of various feature extraction capabilities from the convolution kernels with different sizes. A structural parameter named α is proposed and it searches the optimal combination of convolution kernels with different sizes, and the residual block is introduced and the transposed convolution operation is used instead of the unpooling operation, which improves the final accuracy of height estimation. The experiments carried on the datasets on Guiyang, Huangshan, Geermu demonstrate that the proposed method outperforms the state-of-the-art in mountain areas.

Distributed bundle adjustment method for super large-scale datasets based on LM algorithm

Maoteng ZHENG, Yihui LU, Junfeng ZHU, Xiaoru ZENG, Huanbin QIU, Yuyao JIANG, Xingyue LU, Hao QU, Nengcheng CHEN

2025, 54(5):  899-910.  doi:10.11947/j.AGCS.2025.20240142

Asbtract ( )   HTML ( )   PDF (12170KB) ( )  

Figures and Tables | References | Related Articles | Metrics

This paper proposes a distributed bundle adjustment method based on Levenberg Marquardt (LM) model for large-scale dataset. In order to solve the storage and solution of large-scale coefficient matrix of normal equation, a block-based sparse matrix compression format (BSMC) is used to compress the coefficient matrix of normal equation by taking advantage of its sparse block characteristics. This format also supports distributed storage and update of the coefficient matrix of normal equation. Based on the above compression format, a distributed bundle adjustment method adjustment framework based on the strict LM model has been established. By constructing the normal equations in distributed parallel and parallelizing other steps with high computational complexity, the distributed bundle adjustment for large-scale dataset has been achieved. Through comprehensive comparative experiments between the method proposed in this paper and state-of-the-art methods, the preliminary results show that the memory requirement of this method is significantly reduced, and the data processing capacity is significantly improved. For the first time, we have implemented a bundle adjustment based on the LM strict model on a distributed computing system, which includes real dataset with 1.18 million images and simulated dataset with 10 million images (approximately 500 times the state-or-the-art method based on the LM strict model).

Multi-label scene classification method based on fusion of SAR and optical remote sensing images

Yiming ZHAO, Kelin HU, Kelong TU, Yaxian QING, Chao YANG, Kunlun QI, Huayi WU

2025, 54(5):  911-923.  doi:10.11947/j.AGCS.2025.20240281

Asbtract ( )   HTML ( )   PDF (6463KB) ( )  

Figures and Tables | References | Related Articles | Metrics

Deep convolutional neural networks have proven to be one of the most effective methods for scene classification of high-resolution remote sensing images. Most previous studies focus on scene-level classification of single optical remote sensing images and are primarily limited to single-label classification. However, single optical remote sensing images are often constrained by weather conditions, and single-label annotations cannot fully describe complex image contents. Therefore, in this paper, we constructed a multimodal, multi-label scene classification dataset called SEN12-MLRS, using SAR and optical remote sensing images acquired by the European Space Agency in 2020. We proposed a parallel dual attention fusion network (PDANet) for multi-label scene classification. PDANet achieves optical and SAR image feature extraction as well as multi-modal and multilevel feature fusion through two-branch feature extraction, adaptive feature fusion, and multilevel feature fusion. Experimental results demonstrate that PDANet achieves superior performance compared to many state-of-the-art models on the SEN12-MLRS dataset. The effectiveness of the proposed network and its modules is further validated through ablation experiments.

MAFUNet: water body segmentation algorithm for SAR images combining attention mechanisms and active contour loss

Guangao XING, Guanming LU, Bin HAN

2025, 54(5):  924-936.  doi:10.11947/j.AGCS.2025.20240235

Asbtract ( )   HTML ( )   PDF (10274KB) ( )  

Figures and Tables | References | Related Articles | Metrics

With the increasing development of remote sensing technology, synthetic aperture radar (SAR) imagery has become one of the main methods for water body segmentation. Due to the presence of complex interference, water body segmentation in SAR imagery is still a challenging task. To achieve accurate water body detection, we proposed a multi-level attention fusion UNet (MAFUNet) inspired by the effectiveness of UNet in segmenting small targets with weak edges. First, the spatial attention module (SAM) and channel attention module (CAM) are added to the skip connections between the encoder and decoder parts to extract useful low-level and high-level features, compensating for the loss of semantic information of downsampling. Second, considering the feature distortion resulting from upsampling, an attention upsampler (AU) is designed that retains more detailed information of the image and reduces the effect of introduced noise. Third, the multiscale convolutional pooling block (MCPB) is introduced into the decoder part to better utilize the contextual information, capturing water and land features at different scales. Moreover, an active contour loss is designed as an additional regularization, and a multilayer loss function is used to optimize the network for better extraction of layer-level features, improving the model's segmentation performance. The experimental results show that the proposed MAFUNet outperforms other state-of-the-art models, IoU and F₁ are 94.28% and 97.67% on the ALOS PALSAR dataset.

Cartography and Geoinformation

Methodology for mining causal patterns of multiple geographic elements by considering spatial neighborhood effects

Yan SHI, Shiyi LI, Da WANG, Min DENG, Zhong'an TANG

2025, 54(5):  937-949.  doi:10.11947/j.AGCS.2025.20240369

Asbtract ( )   HTML ( )   PDF (3458KB) ( )  

Figures and Tables | References | Related Articles | Metrics

Geospatial data mining aims to deeply reveal the complex distribution rules and spatio-temporal evolution trends of multiple geographic elements. Current geospatial data mining studies were mostly based on the assumption of spatial correlation dependence, which lacked the analysis of underlying spatial causalities, so the mixed pseudo-correlations would lead to biased or even erroneous mining results. In this case, based on the causal inference theory, this study proposed a multivariate geographic element causal pattern mining method by considering the influence of spatial neighborhood effects on causal relationships. First, the transaction set adapted to the distribution density of geographic elements is automatically created using the spatial clustering algorithm. Then, the spatial causal directed graph structure of multiple geographic elements is constructed by integrating the spatial neighborhood effects and the Bayesian network modeling idea. Finally, the backdoor criterion is used to implement the intervention operation to realize the quantitative calculation of causal effects among multiple geographic elements. In the experiments, the spatial distribution data of urban facilities in Shenzhen and Shanghai are utilized for case studies. The comparison results with the spatial correlation pattern mining method show that the proposed method eliminates the spurious spatial correlations caused by confounding variables, and can effectively obtain the directed causal relationships and causal strength between different types of urban functional facilities, and reveal the local aggregation effects of urban functional facilities more accurately. The mined causal patterns have the potential of providing more credible decision supports for the optimization of urban space layout.

Constructing grade-separated junctions based on combination of local and long-term trajectory feature

Fengwei JIAO, Longgang XIANG, Yuanyuan DENG, Xin CHEN, Huayi WU

2025, 54(5):  950-962.  doi:10.11947/j.AGCS.2025.20240130

Asbtract ( )   HTML ( )   PDF (4383KB) ( )  

Figures and Tables | References | Related Articles | Metrics

The grade-separated junction is a complex multi-level structure composed of roads intersecting longitudinally and transversely. With the modernization of transportation, extracting the geometric and topological structures inside the grade separation from trajectory data is a key step to build a refined navigable road network. Most existing methods extract roads based on the local distance and orientation similarity of trajectory units, which may lead to errors such as merging nearby roads and missing structures. Therefore, this paper proposes a method to construct the road network of highway interchanges based on combination of local and long-term trajectory feature. Firstly, based on the continuity of trajectories, trajectory segments are classified into trajectory clusters with unified paths. The central line of the trajectory clusters is extracted using an adaptive binarization method. Then, considering the local direction feature and long-term distance feature, the central line tracking from the same entrance is carried out to dynamically capture the diversion phenomenon of nearby parallel roads and extract the subnetwork. Finally, the node candidate set of the subnetwork is clustered and the road segments are truncated based on node information. All the subnetworks are merged to generate the road network structure. The road network construction experiments using crowdsourcing trajectory data in Shenzhen are conducted to evaluate the performance of the proposed method, and results show the effectiveness and high accuracy in geometric and topological information. The overall GEO-F₁ score reaches 94.5%, and the TOPO-F₁ score is 94.8%, which outperforms existing state-of-the-art methods.

Summary of PhD Thesis

Research on geospatial grid region name model and experimental system

Daoye ZHU

2025, 54(5):  963-963.  doi:10.11947/j.AGCS.2025.20230555

Asbtract ( )   HTML ( )   PDF (864KB) ( )  

References | Related Articles | Metrics

Research on multi-source data fusion technology and its application in mining subsidence monitoring

Rui WANG

2025, 54(5):  964-964.  doi:10.11947/j.AGCS.2025.20230556

Asbtract ( )   HTML ( )   PDF (867KB) ( )  

References | Related Articles | Metrics

The inversion of terrestrial water storage anomaly based on deep-learning algorithm and fusion of GNSS and GRACE

Yifan SHEN

2025, 54(5):  965-965.  doi:10.11947/j.AGCS.2025.20230564

Asbtract ( )   HTML ( )   PDF (873KB) ( )  

References | Related Articles | Metrics

The retrieval method of GNSS-R sea surface height based on deep learning

Qiang WANG

2025, 54(5):  966-966.  doi:10.11947/j.AGCS.2025.20230566

Asbtract ( )   HTML ( )   PDF (871KB) ( )  

References | Related Articles | Metrics