Due to the coupling effects of climatic conditions, surface and subsurface conditions, and human activities, soil moisture is highly heterogeneous on spatial and temporal scales. The SMAP soil moisture products from satellite microwave remote sensing can be used from continental to global scales, but they are not suitable for small- and medium-scale applications due to low spatial resolution. To improve the spatial resolution of soil moisture products, various downscaling methods have been developed, with the empirical downscaling method being widely used due to its relatively simple calculation. These models require downscaling factors, which are mostly obtained based on optical remote sensing and are susceptible to cloud influence. Therefore, it is impossible to obtain high spatial resolution soil moisture continuously over time using this model for downscaling. To solve this problem, we proposed a downscaling framework based on multiple data sources using machine learning and deep learning methods. The main idea is to use traditional machine learning methods in the absence of clouds and super-resolution methods to downscale soil moisture in the presence of clouds. The combination of these two methods yields time-continuous, high-resolution soil moisture. First, multi-source data were used to obtain fifteen downscaling factors, including surface temperature, normalized vegetation index, albedo, elevation, slope, slope direction, soil cover type, soil texture, etc. Then, three machine learning models (Random Forest, LightGBM, and XGBoost) were used to establish empirical downscaling models of SMAP soil moisture product data with downscaling factors. The best performing XGBoost model was chosen to downscale the spatial resolution of SMAP soil moisture products from 9 km to 1 km. Finally, the DSCGAN super-resolution model was trained based on 9 km and 1 km soil moisture data pairs. The trained models were used to obtain spatio-temporally continuous soil moisture data for the study area. The results show that, by comparing the downscaled soil moisture and original SMAP data, the R is 0.96, the RMSE is 0.034 m³/m³, the bias is 0.011 m³/m³, and the ubRMSE is 0.034 m³/m³. The comparison between the downscaled soil moisture and the measured data shows that the R is 0.696, the RMSE is 0.192 m³/m, the bias is -0.171 m³/m³, and the ubRMSE is 0.089 m³/m³. The downscaling method proposed in this study provides a framework for generating higher resolution spatio-temporally continuous surface soil moisture that can meet the needs of small-scale applications such as regional moisture surveys and agricultural drought monitoring.

Construction waste is an inevitable byproduct of urban renewal processes, causing serious environmental pollution and ecological pressure. Precisely quantifying the annual production of urban construction waste and the resource conversion rate is crucial for assessing the cost of urban renewal. Traditional manual methods of estimating construction waste production rely heavily on statistical data and historical experience, which are inflexible, time-consuming, and labor-intensive in practical application, and need improvement in terms of accuracy and timeliness. Existing deep learning models have relatively poor capabilities in extracting and integrating small targets and multi-scale features, making it difficult to handle irregular shapes and fragmented detection areas. This paper proposes a Multi-Scale Feature Fusion and Attention-Enhanced Network (MS-FF-AENet) based on High-resolution Remote Sensing Images (HRSIs) to dynamically track and detect changes in buildings and construction waste disposal sites. This paper introduces a novel encoder-decoder structure, utilizing ResNet-101 to extract deeper features to enhance classification accuracy and effectively mitigate the gradient vanishing problem caused by increasing the depth of convolutional neural networks. The Depthwise Separable-Atrous Spatial Pyramid Pooling (DS-ASPP) with different dilation rates is constructed to address insufficient receptive fields, resolving the issue of discontinuous holes when extracting large targets. The Dual Attention Mechanism Module (DAMM) is employed to better preserve spatial details, enriching feature extraction. In the decoder, Multi-Scale Feature Fusion (MS-FF) is utilized to capture contextual information, integrating shallow and intermediate features of the backbone network, thereby enhancing extraction capabilities in complex scenes. MS-FF-AENet is employed to extract and analyze changes in building areas at different time periods, calculating the engineering waste from new constructions and demolition waste from demolished buildings, thereby obtaining the annual production of urban construction waste. Furthermore, MS-FF-AENet is utilized to extract construction waste disposal sites at different time periods, estimating the amount of construction waste landfill based on changes in landfill waste, indirectly assessing the resource conversion rate of urban construction waste. Based on HRSIs of Changping District, Beijing from 2019 to 2020, experimental results demonstrate: (1) Among a series of baseline models including UNet, SegNet, PSPNet, DeepLabV3+, DSAT-Net、ConvLSR-Net and SDSC-UNet, MS-FF-AENet exhibits advantages in terms of precision and efficiency in extracting buildings and construction waste; (2) During the period from 2019 to 2020, the annual production of construction waste in the study area due to urban renewal is approximately 4 101 156.5 tons, with approximately 2 251 855.872 tons being landfill waste and approximately 1 849, 300.628 tons being resource conversion waste, resulting in a construction waste resource conversion rate of 45.09%, further corroborating government statistical reports. This paper provides a convenient and effective analysis approach for accurate measurement of the cost of urban renewal.

The precision of individual tree segmentation is important for survey of forest resources. However, traditional individual tree segmentation algorithms suffer from issues such as near tree confusion and low computational efficiency when processing large-scale point cloud data. To address these issues, this paper introduces an improved K-means clustering method that combines spectral clustering and particle swarm optimization for individual tree segmentation of airborne LiDAR point clouds. The proposed method is designed to overcome the limitations of conventional methods by increasing the accuracy of tree segmentation and optimizing the processing of large and complex point cloud data. By combining advanced techniques in spectral clustering and particle swarm optimization, the proposed method significantly improves the precision and efficiency of individual tree segmentation. Firstly, the voxelization of the point cloud data is performed using the Mean Shift algorithm, where the adaptive bandwidth and Gaussian kernel function compute the similarity between voxels, resulting a Gaussian similarity graph reflecting the properties of voxels. This graph not only encapsulates the space structure of the forest but also improves the accuracy of the subsequent data analysis and processing. After voxelization, the Nyström method is applied to efficiently manage the Gaussian similarity graph. This method uses K-nearest neighbor search to select representative samples, effectively reducing the computational burden associated with spectral clustering when dealing with large-scale datasets. By selecting representative samples, the algorithm ensures that the main features of the data are retained, facilitating a more manageable and accurate clustering process. This method optimizes the processing of large amounts of point cloud data by balancing computational efficiency with the requirement to maintain data integrity and accuracy, thus providing a robust foundation for accurate tree segmentation. Using the Nyström approximation, approximate eigenvectors of the similarity graph are obtained, facilitating an effective mapping from the high-dimensional space to a low-dimensional feature space. Finally, the particle swarm optimization algorithm is introduced to enhance the K-means clustering process. This optimization algorithm first randomly initializes a set of particles, each representing a set of potential cluster centers. In each iteration, the particles update their clustering speed and position based on the best historical position of the individual and the best historical position of the group, adjusting the clustering centers to minimize the internal cluster distance. In this paper, publicly available point cloud data from NEWFOR is selected for experiments. The experimental results show that the segmentation results obtained by the proposed algorithm are 5.3% higher in accuracy and 23 times more efficient than the comparison algorithm.

Urban 3D modeling is indispensable for digital twinning and the development of smart cities. The effective extraction of building outlines is a critical step in achieving high-precision urban modeling and 3D mapping. At present, the extraction of building outlines from airborne point cloud data still faces challenges, such as low efficiency and accuracy with conventional methods and limited calibration samples. In response to these challenges, this paper introduces a deep learning method for extracting building outlines from 3D airborne point clouds. The airborne LiDAR point clouds are the primary data input. First, through vertical projection to the XOY plane, point clouds of buildings with the application of progressive morphological filtering are converted to rasterized elevation that characterizes spatial variation of terrain and visible light raster images that depict texture differences. Then, the deep learning model based on Lines-Convolutional Neural Networks (Line-CNN) is employed to preliminarily extract line features from raster images, encompassing stages of feature extraction, node prediction, route generation, and others. To enhance the quality of the primary straight-line extraction, an optimization strategy is introduced, which incorporates a range of comprehensive trimming and completion operations, aligning with information extracted from both the elevation and visible light raster images. Simultaneously, false line segments are eliminated, and missing lines are added, resulting in the regular and complete building outline features. To verify the proposed model, the airborne point cloud data from NUIST campus and the ISPRS H3D 2019 datasets are utilized in the experiment. Our results show that the proposed method accurately and comprehensively extracts building outline features from LiDAR images, achieving an impressive average accuracy and completeness rate, both up to 90%. Furthermore, the proposed method is highly efficient and effectively addresses the challenge of insufficient 3D calibration samples in traditional methods, making it suitable for various applications, particularly large-scale urban 3D modeling and cadastral surveying. To sum up, the proposed method constitutes a significant stride in advancing urban modeling and 3D mapping. It provides a novel solution to address the challenges associated with building outline extraction, particularly within the context of smart cities and digital twins. Due to the model's high accuracy, completeness, and efficiency, our method is highly helpful for a wide range of applications in the urban planning and geospatial information fields.

In the context of rural revitalization, it is important to explore the layout influence factors of rural express delivery points for rural infrastructure construction. In this paper, we propose the Deformed Thiessen Polygon theory based on the current development status of the Thiessen Polygon theory and the needs for the socio-economic application research. Using the Deformed Thiessen Polygon, this study considers the rural area in Suzhou City, Jiangsu Province (its meaning is explained in the main text) as the study area, with the express delivery points as the specific case. Methods such as Ordinary Least Squares (OLS), Spatial Lag Model (SLM), Spatial Error Model (SEM), and Geographically Weighted Regression (GWR) were used to reveal the layout influence factors of rural express delivery points. The main conclusions drawn from this paper are as follows. (1) The Deformed Thiessen Polygon is an improvement and expansion of the Thiessen Polygon for socio-economic applications. It consists of two parts: the deformation area and the non-deformation area. The deformation area represents a local deformation of the Thiessen Polygon, while the non-deformation area retains the original Thiessen Polygon. The Deformed Thiessen Polygon can effectively solve the mismatch problem between the spatial division of the Thiessen Polygon and actual situations, improving the feasibility and scientificity of social and economic research. (2) Using the service scope as a characterization indicator of express delivery point layout, with this indicator as the dependent variable and population size, economic size, road network density, etc., as independent variables, we explored the layout influence factors of express delivery points in the study area. When the probability is less than 0.05, the OLS and SLM results show that population size, average slope, and water area are positive influence factors, with population size being the dominant positive influence factor; Economic size is a negative influence factor. The GWR results show that population size and water area have positive spatial effects, while economic size has a negative spatial effect. There are both positive and negative spatial effects on the road network density and average slope. Based on the analysis presented throughout this paper, the main considerations for the development of the rural express delivery industry are as follows. (1) Give full play to the basic guarantee role of public express delivery points. (2) Emphasize the decisive role of population and economic factors. (3) Guide the comprehensive layout for express delivery points. (4) Create relatively equal rural express delivery services. The Deformed Thiessen Polygon helps to advance the development of the Thiessen Polygon, providing a new basis for spatial division in future studies of socio-economic applications. At the same time, the results of this paper have value for understanding the layout influence factors for rural express delivery points and for implementing rural revitalization strategies.

The statistical analysis methods of spatial data have also been introduced into economic literature after the new economic geography theory brought spatial elements into the research scope of mainstream economics. The agglomeration economy, with geographical proximity as its basic feature, has become an important research object and has gradually become the cross-research field of economic geography and GIS. The Ripley's K function belongs to the distance-based method in spatial point pattern analysis, also known as multi-distance spatial cluster analysis. The DO index develops on this basis. Compared with the K function, the DO index is more in line with the calculation standard of incomplete random distribution of enterprises in economics. In 2014, the coordinated development of the Beijing-Tianjin-Hebei region became a major national strategy. To verify the effectiveness of policy implementation, based on the economic census data of 2008 and 2018, we cleaned the data with the China Economic Census Yearbook as the reference standard. Then, we used the DO index to measure the characteristics and dynamic evolution process of manufacturing co-agglomeration before and after the coordinated development of the Beijing-Tianjin-Hebei region strategy was proposed. The results show that: (1) In 2008 and 2018, 85.47% and 82.53% of manufacturing industry pairs respectively experienced co-agglomeration, with the average intensity changing from 0.030 to 0.029. The basic pattern of co-agglomeration remained stable over the past 10 years. However, the TOP 20 industry pairs with high co-agglomeration intensity value have shown the characteristics of obvious diffusion to Hebei Province. (2) The scale of co-agglomeration has expanded. In 2008, the significant scale of co-agglomeration for all industry pairs was 25-68 kilometers, while in 2018, it was within 55 kilometers and 75~103 kilometers. The average scale of co-agglomeration for the TOP 20 industry pairs increased from 80.91 kilometers to 125.15 kilometers. The critical value of the provincial and municipal boundaries is about 75 kilometers. After the proposal of the Beijing-Tianjin-Hebei coordinated development strategy, there has been more cooperation between manufacturing industries across provincial administrative boundaries, and this characteristic is more evident in the TOP 20 industry pairs. (3) The co-agglomeration between manufacturing industries exhibits a "polarization" characteristic. Industry pairs with high values of co-agglomeration intensity are higher, but those with medium intensity of co-agglomeration have become lower. The scale of the first local high value in 2018 was 54.5 km, smaller than 68 km in 2008, but the overall significant scale was larger than in 2008. This indicates that those with high values are higher and those with low values are lower, which to some extent verifies that there is a circular cumulative causal effect of co-agglomeration. The research results show that the coordinated development of industries in the Beijing-Tianjin-Hebei region has achieved initial success, and relieving Beijing of functions non-essential to its role as China's capital is an important driving force for the coordinated development of Beijing-Tianjin-Hebei industries at this stage.

PM_2.5 concentration prediction plays a pivotal role in the prevention and control of air pollution. Traditional forecasting models, such as the Graph Convolutional Network (GCN) and other spatial-temporal prediction models, measure the spatial correlation of PM_2.5 distribution primarily by monitoring the Euclidean distance between monitoring stations. However, these models often fail to account for the anisotropy effects of terrain, wind direction, and other factors that significantly influence the transport process of air pollutants. This oversight can result in lower accuracy of prediction results, especially in areas with complex terrain. This paper proposes a novel spatiotemporal convolutional network prediction model for PM_2.5 concentration that takes into account the anisotropy of the geographical environment. The model first constructs the edges of the GCN by incorporating the anisotropic effects of terrain and wind direction on PM_2.5 propagation between stations. It then models the station PM_2.5 concentration, land use, and other meteorological factors as node characteristics of the GCN. The model employs GCN to extract the spatial characteristics of PM_2.5 concentration and subsequently uses a Gate Recurrent Unit (GRU) to model and predict the temporal characteristics of PM_2.5 concentration at the station. The model's performance was evaluated using hourly PM_2.5 concentration records from the mountainous Guizhou province in 2017. It was compared against several spatiotemporal prediction baseline models, including Geographically and Temporally Weighted Regression (GTWR), Spatio-Temporal Support Vector Regression (STSVR), and a combined GCN+GRU model. The experimental results demonstrate that the model proposed in this paper significantly outperforms the baseline models. The Root Mean Square Error (RMSE) and Mean Absolute Error (MAE) ratios of PM_2.5 concentration predicted by the model are 10.047 and 6.848, respectively, which represent decreases of 11.29% and 12.16% compared to the baseline models. The R-squared value of 0.883 indicates an average improvement of 3.72% from the baseline. Furthermore, the analysis of the influence of different terrains on the correlation of PM_2.5 concentration at different stations reveals that mountainous, gully, and other terrain features can significantly affect the correlation of PM_2.5 concentration between stations. This, in turn, impacts the prediction results of PM_2.5 concentration at different stations. The study concludes that fully considering the anisotropic effects of terrain and wind direction on PM_2.5 propagation can substantially enhance the prediction accuracy of PM_2.5 in mountain and gully terrain areas. By integrating the anisotropic characteristics of the geographical environment into the prediction model, this research contributes to the development of more accurate forecasting tools for PM_2.5 concentrations. This development is expected to ensure accurate prediction of PM_2.5 concentrations in areas with complex terrain and support the development of effective air pollution control strategies.

As one of the main components of air pollutants, PM_2.5 seriously affects human health, causing issues such as respiratory system damage and increased incidence of cancer and cardiovascular diseases. A complete spatio-temporal dataset of PM_2.5 is key to realizing air pollution control. However, the current PMPM_2.5 datasets often have missing values due to machine failures, human errors, atmospheric conditions, and other factors. Addressing the problems that existing methods for reconstructing missing values fail to fully consider daily periodicity, spatial-temporal heterogeneity of PM_2.5, and the complex nonlinear relationships with the influencing factors, this paper proposes a Daily Periodicity-based Spatial-Temporal Interpolation method (DP-STF) to reconstruct the missing values of PM_2.5 measurements. The method first uses the daily observation data as the processing unit to screen for the optimal spatial stations and time series based on spatio-temporal correlation for the missing locations in both temporal and spatial dimensions. It then utilizes the P-BSHADE (Point Estimation Model of Biased Sentinel Hospital-based Area Disease Estimation) method to reconstruct the missing values of PM_2.5 stations, considering temporal and spatial periodicity. The approach iteratively takes into account spatio-temporal heterogeneity in the initial estimation of the missing data. Finally, Stacking integrated machine learning is used to fit the complex spatio-temporal nonlinear relationships of PM_2.5. The initial spatio-temporal estimates and the PM_2.5 impact factor drive the training of the Stacking integrated model, which is then used for estimating the missing PM_2.5 data. Using the hourly-scale PM_2.5 station data of Beijing-Tianjin-Hebei from 2020 as the research object, the missing data are reconstructed and compared with seven classical methods using the DP-STF method. The experimental results show that, compared to the classical methods, DP-STF achieves superior accuracy. The average RMSE and MAE of this method are reduced by at least 39.83% and 40.12%, respectively, and the R² is improved by at least 5.56%. Additionally, this method effectively captures the extreme values of PM_2.5, significantly increasing the prediction accuracy of the model in spatio-temporal non-stationary regions.

The recognition of building group pattern is crucial for building integration, and an efficient building group pattern recognition method can significantly enhance the quality of automatic map synthesis. Geometric and machine learning/deep learning approaches are the two main strategies that have been most frequently utilized in the field of building group pattern detection in recent years. However, there are limitations associated with geometric approaches, such as challenges in threshold setting, complex rule formulations, and limited pattern recognition capabilities. Techniques based on machine learning and deep learning also have difficulties, such as substantial data requirements and complicated feature selection procedures. In response to these challenges, researchers have developed the directional entropy as a novel approach for multi-pattern detection of building groups. The directional entropy is a derivative measure of information entropy, which has been utilized in spatial analysis to evaluate the uncertainty of directional random variables. It assists in describing the prevalence, characteristics, and regularities of spatial phenomena. The study procedure for utilizing directional entropy in developing group pattern recognition is as follows: First, a minimal spanning tree geometric model is created by clustering building data from Lanzhou City using an artificial visual technique; Then, the building groups are split into sample set 1 and sample set 2, in a 7:3 ratio. Classification thresholds for straight, grid, and irregular building groups are calculated based on the training set and validated using the validation set. The experimental results show that directional entropy achieves a classification accuracy of above 97% for all three different building group types. The classification criteria established on the training set are further applied to building data from Shanghai, which yields expected results. These results demonstrate the effectiveness of directional entropy in classifying various building group modes and highlights the potential of directional entropy in identifying building group patterns. Compared to conventional and machine learning techniques, directional entropy overcomes several limitations and produces satisfactory classification results, presenting a novel strategy and technique for establishing group pattern recognition.

In the context of ecological civilization construction, the comprehensive implementation of the River Chief System has become a powerful measure for China to address complex water problems and maintain the ecological health of rivers and lakes. However, there are still some issues, such as insufficient theoretical research and a lack of design methods for river management maps based on the River Chief System. To address these issues, we have explored the following aspects of research. Firstly, this study interpreted the connotation of the River Chief System from four aspects: organizational structure, river chief responsibilities, main tasks, and supervision and assessment. From the perspective of maps, we analyzed the information transmission and mapping requirements of relevant departments and river chiefs at all levels. Based on the thematic map theory, the concept of a river management map is proposed, and its characteristics were summarized. Building on this, we proposed the design concept of "multi-department collaborative linkage and multi-level river chief task decomposition" from both horizontal and vertical coordination dimensions, thereby establishing a content framework for river management maps. Secondly, based on the theory of geographic scenarios, this study summarized and analyzed the elements of time, place, people, things, events, and phenomena in river management scenarios. We proposed that river management maps should select appropriate spatiotemporal scales, representation methods, visual variables, user interactions, and communication media for user groups to guide the selection of mapping expression content for each map group. The presentation forms of information in the river management dashboard map mainly include maps, information charts, and forms, which were also classified and organized in this study. Then, based on the application scenario of the map, this study adopted an interactive dashboard format and electronic screens as the communication medium, utilizing the five elements user experience model and drawing upon theories of cognition, visualization, and map interaction to design the information architecture, interface layout, color scheme, information presentation, and interactive features of large-screen maps. Finally, taking Baoshan District, Shanghai, as a case study, a river management dashboard map system is designed and implemented, followed by a systematic usability evaluation. The results indicate that the dashboard map supports the display of multi-scale river management information, facilitating clear visibility of target information for users. The aim of this study is to actively contribute to the standardization and regularization of river management maps tailored to the River Chief System, facilitating multi-departmental collaborative management and promoting co-construction and resource sharing.

Visibility refers to the spatial extent that an observer can see from a specific location under certain environmental conditions. Existing visibility analysis methods, based on point cloud data, mainly focus on the outdoor perspective and do not fully consider the problem of point cloud voids. To fill the gaps in the current indoor visibility research, this paper proposes an indoor 3D visibility analysis method based on point cloud data from UAVs and ground survey data. Our method reduces the influence of point cloud voids and aims to improve the accuracy of the visibility analysis. The method presented in this paper is divided into three parts: data preprocessing, visual space construction, and visual distance correction. In the data preprocessing stage, UAV point cloud data is used to extract the building roof contour and DEM, which are utilized for subsequent visual distance correction. Ground-based point cloud data is then used to extract the center points of the windows on the building façade, which serve as observation viewpoints. In the viewspace construction stage, the viewpoint is taken as the center, and the theoretical visual distance is taken as the radius. Points within the theoretical visual distance are selected to construct a hemispherical viewspace. The points in the visual field space are projected onto the depth image to form the visual field image. The visual field of view is then calculated, and the visual range of the field of view space is analyzed to simulate the external scene observed through the window in the room. During this process, point cloud voids can lead to errors in the visual range. Specifically, when the positions of the scanning center and the viewpoint differ significantly, missing parts may appear in the visible space, and the pass-through errors may also occur near the viewpoint due to insufficient point cloud density. In the sight distance correction stage, this paper categorizes the viewspace using spatial attributes and proposes a sight distance correction method to attenuate the effect of point cloud voids on the volume of the visible space by considering the feature continuity and occlusion relationship within different categories of the viewspace. Finally, visibility analysis is conducted using the visual space volume index at the viewpoint and is compared with the existing voxel-based and surface-based visibility analysis methods, as well as a manual evaluation. The DTW distances for comparisons are 48 247, and 240, respectively. The results show that the proposed method has better consistency with manual evaluation results. It can analyze the visibility of indoor environments more effectively, and is also applicable to outdoor visibility analysis. This method provides a comprehensive and reliable visibility analysis strategy for urban planning, landscape design, and other related fields.

Facility location is of great significance for improving residents’ quality of life, and geographic accessibility indicators, such as the road network, are often used as the main decision-making factors. Clustering analysis based on geographic accessibility is an important tool for solving such problems. However, existing clustering algorithms often fail to guarantee the accuracy of clustering results, the accessibility of cluster centers, or the selectivity of cluster centers, making them less effective in solving the facility location problem in real scenarios. This paper proposes a Fuzzy C-Means clustering algorithm based on Reachable Distance (FCM-RD), which modifies the objective function, the membership function, and the cluster center function of the classical FCM. It employs reachable distance as a measure of geographic reachable similarity and iterates the cluster centers during the clustering process. Specifically, to capture the true relationships and connectivity between different elements, FCM-RD takes into account physical and spatial barriers, employs the shortest path distance along the road network as the reachable distance, and aligns geographic coordinates with the road network. It is possible for one position on the road network to correspond to multiple positions in geographic coordinates. Consequently, when multiple candidate positions for cluster centers are obtained, a cluster center correction mechanism is designed to iterate the accessible cluster center with reachable distance during the clustering process. Mathematical analysis and experiments in actual scenarios both show the validity of the cluster center iteration mechanism, showing the selected cluster centers in each iteration of FCM-RD are the unique and minimum value points of the intra-cluster objective function. The rationality of FCM-RD is further verified through experiments, and it is compared with baseline algorithms from three aspects: experimental results, convergence, and performance. The results indicate that, compared to the baseline algorithms, FCM- RD improves performance on both the mean and maximum indicators of the shortest reachable distance, with some indicators even improving by up to 38.9%. In a few experiments, there are slight improvements in the DB index and silhouette coefficient indicators, and 100% of the cluster centers selected by FCM- RD are located on the road network. FCM- RD overcomes the shortcomings of ignoring geographical obstacles and unreachable cluster centers. In conclusion, FCM-RD not only obtains accessible cluster centers without location restrictions but also achieves better clustering results. FCM-RD provides an effective and precise solution for geographical spatial clustering in practical scenarios.

The development of intelligent connected cars and autonomous driving has become an important national strategic direction of our country. As the core infrastructure of autonomous driving, high-precision maps have high commercial value, and their safety issues cannot be ignored. Based on invisible characters, this paper presents a digital watermarking algorithm for high precision maps in OpenDrive format. First, a mapping mechanism between watermark information and invisible characters is constructed, and a composite watermark character sequence containing watermark bit value, bit position information, and Hamming code is established based on the original watermark information, which ensures the synchronization of watermark bits and watermark index while providing the algorithm with a certain self-correction ability. Then, the mapped invisible characters are embedded into the high-precision map attribute values in the order of the pseudo-randomly scrambled identifier set and its attribute values, so that the watermark information is embedded into the data irregularly and distributed evenly. When detecting the watermark, the watermark bits and its corresponding position information are extracted according to the embedding sequence. The final watermark information is obtained according to the majority voting mechanism, and the error bits are corrected according to the Hamming code. Experiments show that the algorithm can fully resist the attacks of translation, rotation, and scaling while maintaining the map accuracy, and exhibits strong robustness to clipping attacks, which can be applied to the copyright protection of high precision maps in OpenDrive format.

The essence of geospatial knowledge lies in unveiling the spatiotemporal distribution, dynamics of change, and interaction patterns of geographical entities and phenomena. However, existing knowledge base management platforms often overlook the specific needs of geospatial knowledge representation and lack the capability to handle the unique attributes of geospatial data, making it challenging to meet the requirements for constructing and applying geospatial knowledge graphs. The Geospatial Knowledge Base Management System (GeoKGMS) is designed on the basis of an integrated geospatial knowledge base engine that efficiently aggregates geospatial knowledge resources across various modalities—'Image-Text-Number'—automates the construction of geospatial knowledge graphs, and facilitates a one-stop geospatial knowledge engineering process. This paper elucidates four key technologies for managing geospatial knowledge bases. First, the cloud-native geospatial knowledge base microservice unified scheduling technology decomposes the large geospatial knowledge base management system into fine-grained, independently operable, and deployable microservices. By comprehensively managing the lifecycle of the geospatial knowledge base, service classification and orchestration methods are determined to achieve unified scheduling of these microservices. Second, a human-computer collaborative geospatial knowledge graph construction method is proposed, supporting the sustainable, collaborative construction of geospatial knowledge graph engineering. Third, the spatiotemporal hybrid encoding technology of the geospatial knowledge graph achieves unified representation of geospatial knowledge by integrating multimodal geospatial data and spatiotemporal information. Fourth, a multimodal geospatial knowledge integrated storage and large-scale spatiotemporal graph partitioning technology is proposed to address the challenges of efficiently managing complex structured geospatial knowledge and retrieving large-scale spatiotemporal knowledge tuples. Based on these key technologies, an application service framework for GeoKGMS has been designed, featuring six functional modules: geospatial knowledge base management, multimodal geospatial knowledge extraction, human-computer collaborative construction of geospatial knowledge graphs, geospatial knowledge reasoning, geospatial knowledge graph quality assessment, and geospatial knowledge visualization. To demonstrate GeoKGMS's capabilities, the Karst landform knowledge graph is used as a case study. The Karst landform knowledge graph is an integrated 'Image-Text-Number' geospatial knowledge graph, constructed based on geospatial knowledge extracted from the texts, schematic diagrams, and related maps in geomorphology textbooks. Through a collaborative pipeline, geomorphology experts and computers jointly perform tasks such as mapping, alignment, supplementation, and conflict resolution of geospatial knowledge. This collaboration ultimately leads to the automated construction of the Karst landform knowledge graph by GeoKGMS. The resulting graph is highly consistent with expert knowledge models, ensuring the interpretability of knowledge-driven geocomputation and reasoning in practical applications.