China is one of the countries most severely affected by geological disasters. Researching high-precision and highly reliable methods for monitoring and predicting landslide deformation holds practical significance for disaster prevention and mitigation efforts. Using the massive Outang landslide in the Three Gorges Reservoir Area as a case study, this paper addresses the issue of the atmospheric interference in extracting landslide deformation using time-series InSAR technology. To correct for atmospheric effects, the GACOS model is introduced and validated against GNSS observation data. To address the often-overlooked temporal-spatial analysis before landslide deformation prediction, the Moran index and Hurst index are calculated to analyze the spatiotemporal characteristics of landslide deformation. Recognizing that landslide deformation is influenced not only by historical deformation but also by various external factors, this paper proposes coupling landslide influencing factors with deformation data for prediction. A Long Short-Term Memory (LSTM) model, optimized by Variational Mode Decomposition (VMD) and the Sparrow Search Algorithm (SSA), is employed for the prediction. By decomposing landslide displacement data into trend, periodic, and random components using VMD, the LSTM network structure is constructed. SSA is used to determine the optimal number of hidden units, maximum training periods, and the initial learning rate of the LSTM model. Additionally, methods such as data normalization, regularization, and model evaluation are employed to enhance the performance and stability of the LSTM model. Finally, the model is trained using the influencing factors and decomposed displacement data to predict landslide deformation. The results indicate that: (1) From January 2021 to June 2023, the maximum and minimum deformation rates of the Outang landslide were -72.75 mm/a and 50.74 mm/a, respectively; (2) The deformation in the study area exhibits positive spatial autocorrelation, with the landslide in the settlement area showing a persistent trend; (3) The prediction error of the LSTM model optimized by VMD and SSA is only 0.37 mm, representing an 11.004% accuracy improvement compared to the standard LSTM model. Based on time-series InSAR technology and spatiotemporal analysis results, this paper constructs a high-precision prediction model for landslide deformation, incorporating multiple influencing factors. This model provides a valuable reference for the prevention and control of landslide disasters.

[Significance] As a representative of new-quality productivity, the low-altitude economy is gradually emerging as a new engine for economic growth. This economy is based on the development and utilization of low-altitude airspace resources. While bringing development opportunities to geospatial information technology, it also poses entirely new challenges. [Progress and Analysis] In this paper, we introduce the division of low-altitude airspace resources and highlight typical drone application scenarios in the context of the low-altitude economy. Subsequently, we analyze the broad application prospects of geospatial information technology in key areas of the low-altitude economy, including the refined utilization of airspace resources, the construction of low-altitude environments, the planning, construction, and operation of new air traffic infrastructure, as well as the safe and efficient operation and regulatory oversight of drones. We emphasize that the geospatial information industry will benefit from development opportunities such as the integration and innovation of emerging scientific and technological advancements, growing market demand, policy support, industrial guidance, and industrial upgrading and transformation. [Prospect] Finally, we briefly address the challenges geospatial information technology must overcome to meet the development needs of the low-altitude economy. These include advancements in spatio-temporal dimension elevation, map and location-based services, high-frequency and rapid data acquisition systems, all-time and all-domain capabilities, and ubiquitous intelligent technologies. These areas will also serve as future directions for development and breakthroughs in geospatial information technology.

[Objectives] Forecasting is a key research direction in Geospatial Artificial Intelligence (GeoAI), playing a central role in integrating surveying, mapping, geographic information technologies, and artificial intelligence. It drives intelligent innovation and facilitates the application of spatial intelligence technologies across diverse real-world scenarios. [Progress] This study reviews the historical development of GeoAI-driven spatiotemporal forecasting, providing an overview of prediction models based on statistical learning, deep learning, and generative large models. In addition, it explores the mechanisms of spatiotemporal dependence embedding within these models and decouples general computational operators used for modeling temporal, spatial, and spatiotemporal relationships. [Prospect] The challenges faced by intelligent prediction models include sparse labeled data, lack of explainability, limited generalizability, insufficient model compression and lightweight design, and low model reliability. Furthermore, we discuss and propose four future trends and research directions for advancing geospatial intelligent prediction technologies: a generalized spatial intelligent prediction platform incorporating multiple operators, generative prediction models integrating multimodal knowledge, prior-guided deep learning-based intelligent prediction models, and the expansion of geospatial intelligent prediction models into deep predictive applications for Earth system analysis.

[Objectives] To systematically review recent advancements in text-to-image generation technology driven by large-scale AI models and explore its potential applications in urban and rural planning. [Discussion] This study provides a comprehensive review of the development of text-to-image generation technology from the perspectives of training datasets, model architectures, and evaluation methods, highlighting the key factors contributing to its success. While this technology has achieved remarkable progress in general computer science, its application in urban and rural planning remains constrained by several critical challenges. These include the lack of high-quality domain-specific data, limited controllability and reliability of generated content, and the absence of constraints informed by geoscience expertise. To address these challenges, this paper proposes several research strategies, including domain-specific data augmentation techniques, text-to-image generation models enhanced with spatial information through instruction-based extensions, and locally editable models guided by induced layouts. Furthermore, through multiple case studies, the paper demonstrates the value and potential of text-to-image generation technology in facilitating innovative practices in urban and rural planning and design. [Prospect] With continued technological advancements and interdisciplinary integration, text-to-image generation technology holds promise as a significant driver of innovation in urban and rural planning and design. It is expected to support more efficient and intelligent design practices, paving the way for groundbreaking applications in this field.

[Significance] Data resources have become pivotal in modern production, evolving in close synergy with advancements in artificial intelligence (AI) technologies, which continuously cultivate new, high-quality productive forces. Remote sensing data intelligence has naturally emerged as a result of the rapid expansion of remote sensing big data and AI. This integration significantly enhances the efficiency and accuracy of remote sensing data processing while bolstering the ability to address emergencies and adapt to complex environmental changes. Remote sensing data intelligence represents a transformative approach, leveraging state-of-the-art technological advancements and redefining traditional paradigms of remote sensing information engineering and its applications. [Analysis] This paper delves into the technological background and foundations that have facilitated the emergence of remote sensing data intelligence. The rapid development of technology has provided robust support for remote sensing data intelligence, primarily in three areas: the advent of the big data era in remote sensing, significant advancements in remote sensing data processing capabilities, and the flourishing research on remote sensing large models. Furthermore, a comprehensive technical framework is proposed, outlining the critical elements and methodologies required for implementing remote sensing data intelligence effectively. To demonstrate the practical applications of remote sensing data intelligence, the paper presents a case study on applying these techniques to extract ultra-high-resolution centralized and distributed photovoltaic information in China. [Results] By integrating large models with remote sensing data, the study demonstrates how remote sensing data intelligence enables precise identification and mapping of centralized and distributed photovoltaic installations, offering valuable insights for energy management and planning. The effectiveness of remote sensing data intelligence in addressing challenges associated with large-scale photovoltaic extraction underscores its potential for application in critical fields. [Prospect] Finally, the paper provides an outlook on areas requiring further study in remote sensing data intelligence. It emphasizes that high-quality data serves as the foundation for remote sensing data intelligence and highlights the importance of constructing AI-ready knowledge bases and recognizing the value of small datasets. Developing targeted and efficient algorithms is essential for achieving remote sensing intelligence, making the advancement of practical data intelligence methods an urgent research priority. Furthermore, promoting multi-level services for remote sensing data, information, and knowledge through data intelligence should be prioritized. This research provides a comprehensive technical framework and forward-looking insights for remote sensing data intelligence, offering valuable references for further exploration and implementation in critical fields.

In high-resolution remote sensing images, challenges such as blurred visual features of objects and different spectra for the same object arise. Segmenting similar ground objects and shaded ground objects in a single mode is difficult. Therefore, this paper proposes a remote sensing image segmentation model based on multi-modal feature extraction and hierarchical perception. The proposed model introduces a multi-modal feature extraction module to capture feature information from different modalities. Using the complementary information of IRRG and DSM, accurate pixel positions in the feature map are obtained, improving semantic segmentation of high-resolution remote sensing images. The coordinate attention mechanism fully fuses the features from different modalities to address issues of blurred visual features and different object spectra during image segmentation. The abstract feature extraction module uses MobileNetV3 with dual-path bottleneck blocks as the backbone network, reducing the number of parameters while maintaining model accuracy. The hierarchical perception network is introduced to extract deep abstract features, and the attention mechanism is improved by embedding scene perception of pixels. Leveraging the inherent spatial correlation of ground objects in remote sensing images, efficient and accurate class-level context modeling is achieved, minimizing excessive background noise interference and significantly improving the semantic segmentation performance. In the decoding module, the model uses multi-scale aggregation dual fusion for feature recovery, strengthening the connection between the encoder and the decoder. This combines low-level features with high-level abstract semantic features, enabling effective spatial and detailed feature fusion. Progressive upsampling is used for feature recovery, resolving the issue of blurred visual features and improving segmentation accuracy. Based on high-resolution remote sensing images from the ISPRS Vaihingen and Potsdam datasets, the experimental results demonstrate that MFEHPNet outperforms a series of comparison models, including C3Net, AMM-FuseNet, MMFNet, CMFet, CIMFNet, and EDGNet, across various performance indicators. In the ISPRS Vaihingen and Potsdam datasets, MFEHPNet achieves an overall accuracy of 92.21% and 93.45%, an average intersection ratio of 83.24 % and 83.94 %, and a Kappa coefficient of 0.85. The frequency-weighted intersection ratio is 89.24 % and 90.12%, respectively, significantly improving the semantic segmentation performance of remote sensing images and effectively addressing the issues of blurred feature boundaries and different spectra during segmentation.

In the context of the rapid development of urbanization, the reasonable selection of locations for public service facilities is critical for delivering efficient services and enhancing the quality of urban residents' lives. However, prevailing approaches for allocation of public service facilities often fall short of meeting the demands on their performance and efficiency in complex and large-scale real-world scenarios. To address these issues, this article proposed a novel Graph-Deep-Reinforcement-Learning Facility Location Allocation Model (GDRL-FLAM), coupling a Facility Location Allocation Graph Attention Network (FLA-GAT) with a Deep Reinforcement Learning (DRL) algorithm. This proposed model tackled the location allocation problem for public service facilities based on graph representation and the REINFORCE algorithm. To assess the performance and efficiency of the proposed model, this study conducted experiments based on randomly generated datasets with 20, 50, and 100 points. The experimental results indicated that: (1) For the tests with 20, 50, and 100 points, the GDRL-FLAM model exhibited a significant improvement ranging from 11.79% to 14.49% compared to the Genetic Algorithm (GA) which is one of the commonly used heuristic algorithms for addressing location allocation problems. For the tests with 150 and 200 points, the improvement ranged from 1.52% to 9.35%. Moreover, with the increase in the size of the training set, the model also demonstrated enhanced generalizability on large-scale datasets; (2) The GDRL-FLAM model showed strong transfer learning ability to obtain the location allocation strategies in simple scenarios and adapt them to more complex scenarios; (3) In the case study of Singapore, the GDRL-FLAM model outperformed GA significantly, achieving obvious improvements ranging from 1.01% to 10.75%; (4) In all these abovementioned tests and experiments, the GDRL-FLAM model showed substantial improvement in efficiency compared to GA. In short, this study demonstrated the potential of the proposed GDRL-FLAM model in addressing the location allocation issues for public service facilities, due to its generalization and transfer learning abilities. The proposed GDRL-FLAM could also be adapted to solve other spatial optimization problems. Finally, the article discussed the limitations of the model and outlined potential directions for future research.

[Objectives] The outdoor pedestrian navigation road network is a vital component of maps and a crucial basis for outdoor activity route planning and navigation. It plays a significant role in promoting outdoor travel development and ensuring safety management. However, existing research on road network generation mainly focuses on the construction of urban vehicular navigation networks, with relatively less emphasis on hiking navigation road networks in complex outdoor environments. Moreover, existing methods primarily emphasize the extraction of two-dimensional geometric information of roads, while the reconstruction of real three-dimensional geometric and topological structures remains underdeveloped. [Methods] To address these limitations, this study proposes a method for constructing the three-dimensional outdoor pedestrian navigation road network maps using crowdsourced trajectory data. This approach leverages a road network generation layer and an elevation extraction layer to extract the two-dimensional structure and three-dimensional elevation information of the road network. In the road network generation layer, a trajectory density stratification strategy is adopted to construct the two-dimensional vector road network. In the elevation extraction layer, elevation estimation and optimization are performed to generate an elevation grid raster map, which is then matched with the two-dimensional road network to produce the three-dimensional hiking navigation road network. [Results] To demonstrate the effectiveness of the proposed approach, experiments were conducted using 1 170 outdoor trajectories collected in 2021 from Yuelu Mountain Scenic Area in Changsha through an online outdoor website. The constructed outdoor three-dimensional hiking road network map achieved an average positional offset of 4.201 meters in two-dimensional space and an average elevation estimation error of 7.656 meters. The results demonstrate that the proposed method effectively handles outdoor trajectory data with high noise and varied trajectory density distribution differences, generating high-quality three-dimensional hiking road network maps. [Conclusions] Compared to traditional outdoor two-dimensional road networks, the three-dimensional navigation road networks constructed this study provide more comprehensive and accurate map information, facilitating improved pedestrian path planning and navigation services in complex outdoor environments.

Scientific knowledge of the spatio-temporal evolution processes and formation mechanisms of territorial space in countries along the Central Asia-West Asia Economic Corridor holds significant scientific value and practical importance for supporting the current "Going Global" strategy and the "Belt and Road" initiative. Based on the dominant functions of the territories, the Central Asia-West Asia Economic Corridor is divided into three major types of territorial space: urban and rural construction, agricultural production, and ecological protection. A long-term analysis base map of territorial space from 2002 to 2022 was constructed by integrating multi-source spatio-temporal data. The spatio-temporal cube model was employed to depict the spatio-temporal evolution processes and typical patterns, while the integrated spatial transformation intensity model analyzed the characteristics of spatial structural transformation across three dimensions: scale, location, and intensity. The VIVI-SHAP framework of an interpretable machine learning model was used to analyze the evolution mechanisms, focusing on the importance of driving factors, interaction intensity, and non-linear dependencies. The results show that: (1) Approximately 6.14% of the territorial space in countries along the corridor underwent structural transformation over the past 20 years. The proportion of urban and rural construction space, though small, increased steadily by 0.17%, while agricultural production space decreased by 19.04% overall, with significant structural changes within the ecological protection space. (2) The dynamic interchange between green and other ecological spaces within the ecological protection space is predominant, with a systematic tendency for green ecological space to convert into agricultural production space, while the main source of urban and rural construction space expansion was green ecological space, accounting for 56.36% of the total converted area. 3) The territorial spatial pattern of the corridor is shaped by multiple processes of territorial space transformation, each with different magnitudes, intensities, and driving mechanisms. Natural geographic factors and transportation location factors played decisive roles, while the global influence of population growth and socio-economic development on territorial space structural transformation was less pronounced. This study provides new perspectives and methods to reveal the patterns and mechanisms of changes in land spatial types in the Central Asia-West Asia region. It further provides data support for decision-making departments to formulate reasonable land spatial planning, and demonstrates its application value in achieving greater spatial comprehensive benefits and promoting coordinated regional economic development.

[Objectives] With the application of knowledge graph techniques in the field of Geographical Information Science (GIS), the Geographical Knowledge Graph (GeoKG) has become a key research direction. GeoKGs often lack sufficient geographic knowledge coverage, which can negatively impact downstream applications. Therefore, reasoning techniques are essential for GeoKG to complete missing knowledge, identify inconsistencies, and predict trends in geographic phenomena. Unlike reasoning techniques applied to general knowledge graphs, reasoning on GeoKGs must handle the unique and complex spatial and temporal characteristics of geographic phenomena. This paper comprehensively introduces and summarizes recent advances in GeoKG reasoning. [Analysis] First, it introduces the relevant concepts and problem definitions of GeoKG reasoning. Second, it analyzes the two core tasks of GeoKG reasoning: knowledge completion and prediction. The reasoning model for knowledge completion primarily fills gaps in the graph to ensure knowledge integrity, while the reasoning model for prediction aims to forecast future trends based on existing geographic data. These two models are optimized for different application scenarios, with different focuses in processing geographic data. [Prospect] Finally, the paper explores future development trends in GeoKG reasoning, highlighting areas such as processing complex relationships in spatiotemporal data, reasoning with multi-scale geographic knowledge, fusing multimodal data, and enhancing the interpretability and intelligence of reasoning models. Additionally, the integration of GeoKGs with large-scale pre-trained models is expected to become a key area of focus.

Urban green spaces are critical components of urban ecosystems, playing an irreplaceable role in improving the ecological environment and enhancing quality of life. High-precision identification of urban green spaces is fundamental for urban renewal and optimizing green infrastructure. However, research on the identification and spatial heterogeneity of green spaces in megacities remains relatively limited. This study, taking Xi'an as an example, integrates urban street view images and GF-2 (Gaofen-2) satellite imagery, employing methods such as ISODATA classification, K-Means classification, and convolutional neural networks to achieve multi-dimensional, downscaled, and high-precision identification and analysis of green spaces. The results indicate the following: (1) The K-Means classification method demonstrates significantly higher accuracy (84.5%) compared to the ISODATA classification method (62.4%) and more accurately maps the spatial characteristics and heterogeneity patterns of green spaces. The green space coverage identified by the K-Means method is 0.277 0, which is lower than the 0.360 7 identified by ISODATA. (2) The average Green View Index (GVI) of streets in Xi'an's main urban area is 0.156 0, indicating a generally good level of street greening. However, there is notable polarization across different roads, with 30% of sampling points having a GVI below 0.080 0. Overall, the GVI of higher-grade roads is greater than that of lower-grade roads, following the trend: primary roads > secondary roads > trunk roads > tertiary roads. (3) There is a positive correlation between the GVI of streets and the vegetation coverage in their surrounding areas in Xi'an's main urban area. However, this correlation weakens in certain road sections, reflecting differences between vertical cross-sections and overhead views of the streets. Combining these perspectives provides a more accurate assessment and quantification of urban green spaces. This study provides a reference for green space planning, green infrastructure construction, and smart management in Xi'an, as well as technical guidance for high-precision identification and spatial analysis of urban green spaces in other cities.

Enhancing and sustaining urban competitiveness is contingent upon the presence of urban vitality. Urban planners and managers face increasing pressure to find more accurate and logical ways to manage urban development due to the growing challenges associated with municipal government. This study focuses on the central region of Nanjing. A detailed framework for evaluating urban vitality is proposed from three perspectives, human activities, network interactions, and the physical environment. This framework uses foundational road networks and building footprints from the World Map, Baidu heatmap, Dianping restaurant data, social media check-in data, Baidu Map POI, and innovation data. To create a comprehensive vitality evaluation framework, nine urban vitality indicators were reduced in dimensionality using the real-coded accelerated genetic algorithm based on the Projection Pursuit Model (RAGA-PPM). An analysis was also conducted on the differences with EWM and the spatial distribution patterns of both unidimensional and comprehensive vitality in Nanjing. The conclusions can be divided into three parts. First, the spatial distribution pattern of vitality in Nanjing's central urban area is successfully reflected by the comprehensive evaluation technique based on multi-source big data. The validity of the proposed evaluation system was confirmed by analyzing vitality cluster sample locations. Second, similar spatial features may be seen in Nanjing's unidimensional vitality, revealing a monocentric urban structure, with high-value areas gradually decreasing outward from the Xinjiekou commercial district. Commercial districts and metro stations are the focal points of population activity vitality, with each district exhibiting strong central values and secondary vitality clusters. Urban vitality values decline, with the Xinjiekou commercial area and Nanjing South Station serving as hubs of network interaction vitality. Urban vitality ratings decrease concentrically, with the Xinjiekou commercial area and Nanjing South Station serving as the hubs of network interaction vitality. Physical building vitality is geographically scattered, with high and relatively high values spread across most areas. Third, unidimensional vitality is not unfamiliar to comprehensive vitality. Additional viable centers for vitality were identified, with each district having a vitality hub. Xuanwu, Gulou, Jianye, and Qinhuai districts, which comprise the old city, form the core of Nanjing's vibrancy and serve as significant hubs. Liuhe and Yuhuatai districts have the lowest vitality, while Xuanwu and Qinhuai districts show the highest vitality. Most districts with above-average comprehensive vitality scores are located near transportation hubs, university areas, industrial parks, pedestrian streets, and commercial centers. According to the study, urban designers may benefit from a more thorough and multifaceted understanding of urban vitality patterns.

Due to the imbalanced regional development, data scarcity exists in some regions, which to some extent restricts the progress of spatial prediction research. The introduction of cross-area knowledge transfer offers a valuable method for mitigating the impact of data scarcity in areas with limited samples and for conducting spatial prediction. With technological advancements, spatial prediction methods based on transfer learning and the Third Law of Geography have become mainstream in the fields of computer science and geography. Transfer learning techniques leverage knowledge from a source domain with abundant data to solve related tasks in a target domain with limited data. Meanwhile, the proposal and application of the Third Law of Geography show that by comparing the similarity of geographical environmental variables between sampled regions and unsampled regions (rather than relying solely on traditional spatial distance or quantitative relationships), it is possible to predict target information in unsampled regions using a small amount of sample data. This provides a theoretical basis and methodological reference for selecting the source domain and target domain in cross-regional knowledge transfer. This paper conducts a literature review of cross-regional spatial prediction research based on these two major methods since 2018, focusing on the following key tasks: (1) Comparing and analyzing the basic principles of spatial prediction based on geographical similarity and transfer learning, and identifying differences in their technical procedures; (2) Summarizing the differences in similarity representation indicators and measurement methods between the two approaches; (3) Examining differences in commonly used auxiliary data, spatial analysis units, modeling methods, and evaluation indicators between the two prediction methods; (4) Discussing the challenges and limitations faced by these cross-regional knowledge transfer methods. The study shows that while the technical principles of both methods are basically consistent, they have specific limitations regarding their scope of application, similarity representation and measurement, relevant auxiliary variables, and parameter selection. The research offers useful insights for optimizing and improving these methods, integrating them effectively, innovating cross-regional prediction approaches, and expanding their application fields.

[Objectives] The geographic system is an integrated framework encompassing natural and human phenomena and their interrelationships on the Earth's surface. While Geographic Information Systems (GIS) can digitally process these geographic elements, they face challenges in addressing rapidly changing geographic contexts with complex 3D structures. This is primarily due to the lack of bi-directional interactions between physical and informational spaces, as well as their reliance on predefined rules and historical data. In this paper, we propose the concept of a “Geographic Intelligent Agent” as an advanced form of GIS, which integrates embodied intelligence, self-supervised learning, and multimodal language modeling to improve environmental perception, spatial understanding, and autonomous decision-making. [Methods] The architecture of the geographic intelligent agent consists of three core components: multimodal perception, an intelligent hub, and an action manipulation module. These components collectively acquire comprehensive environmental information through sensor networks, perform complex situatio reasoning using knowledge graphs and generative models, and enable real-time control and multilevel planning of the physical environment. To adapt to differences between virtual and real environments, the geographic intelligent agent is tested using the earth simulator and a test field platform, equipping it with stronger autonomous capabilities in complex and dynamic geographic contexts. [Results] This paper also demonstrates the implementation of geographic intelligent agent in spatial intelligence applications using the virtual digital human “EarthSage” as an example. [Conclusion] As a prototype of the geographic intelligent agent, "EarthSage" integrates modules such as the spatiotemporal Knowledge Ggraph (GeoKG) and a Cognitive Map Generation Model (GeoGPT), assisting users in obtaining intelligent spatial decision-making support in fields such as emergency management, urban planning, and ecological monitoring. This work exemplifies the transformation of GIS from a traditional information processing tool to an autonomous spatial intelligent system, marking a significant advancement in the field.

[Background] Chinese Classical Gardens (CCGs), as integral components of world cultural heritage and essential urban recreational spaces, hold profound cultural, historical, and aesthetic value. Renowned for their intricate design, these gardens provide cultural ecosystem services through dynamic interactions between tourists and landscapes. Visual perception plays a pivotal role in these interactions, directly influencing how visitors engage with and interpret the "scenery"—a concept central to CCGs. With rapid advancements in 3D real scene reconstruction and digital simulation technologies, a pressing challenge has emerged: developing a 3D data model for CCGs tailored to visual perception computing. Traditional models fail to capture the complex interplay between spatial elements and human perceptual responses. [Objectives] This study aims to address this challenge by tackling three core methodological issues: (1) constructing a visual perception framework to represent the unique "scenery" concept inherent to CCGs; (2) analyzing tourist behavior through the lens of visual perception processes; and (3) organizing a 3D data model that supports robust analysis and visualization. [Methods] To systematically address these challenges, the study elaborates on a visual perception framework for CCGs, integrating four critical stages of visitors' visual experiences: object (what is seen), path (how one navigates), subject (who perceives), and outcome (the resulting impressions and emotions). This framework incorporates spatial narratives, consisting of a narrative symbol system and strategies, and landscape space composition, distinguishing among environmental space, visual perception space, and visual cognition space. Building on this framework, a novel 3D data model tailored to visual perception computing in CCGs is proposed. The model is structured into three interrelated layers: the physical features layer (capturing spatial and structural details), the behavior patterns layer (analyzing tourists' movements and gaze behaviors), and the analytical layers (integrating visual perception metrics). [Results] The feasibility of the proposed approach is demonstrated through a case study of the Humble Administrator's Garden in Suzhou. The implementation process involves acquiring physical data, configuring behavioral data, setting up the storage environment, and computing visual perception. This multi-layered approach provides a theoretical framework for understanding visual perception in CCGs and establishes a methodological pathway for applying 3D technologies to cultural heritage research. [Conclusions] The proposed 3D data model offers a deeper understanding of visual perception within CCGs, facilitating new insights into spatial design and visitor experiences. Furthermore, the methods outlined in this paper have broader implications for studying and preserving other cultural heritage sites, advancing the integration of digital technology in heritage conservation and cultural landscape analysis.

Scientific and accurate monitoring of mangroves is the basis and premise for protecting marine coastal wetland ecosystems. Multi-source remote sensing data can be used to classify mangrove species effectively, but challenges remain in applying optical and SAR image features along with their time-varying information. In this paper, based on Sentinel-1/2 image data, we propose a mangrove species classification framework using Multi-source Features-coupled and Ensemble Learning algorithm (MFEL). The framework analyzs the classification advantages of spectral index features, SAR polarization features, and their temporal harmonic spectral features in feature selection and coupling. It then stacks the Random Forest (RF) and eXtreme Gradient Boosting (XGBoost) models to construct an Ensemble Learning model for mangrove species classification. Comparing the RF classification model and XGBoost models based on feature optimization, we evaluated the classification accuracy and feature application differences of the MFEL classification method. Zhanjiang Mangrove Forest National Nature Reserve was selected as the experimental area. The results show that: ① compared to using only spectral index features, classification accuracy improves by 6% and 8% with the addition of SAR polarization features or temporal harmonic spectral features, respectively. Adding both SAR polarization features and temporal harmonic spectral features simultaneously improves classification accuracy by 12%, making it more effective for mangrove species classification. ② The MFEL method achieves the highest classification accuracy, with an overall accuracy of 88.03% and a Kappa coefficient of 0.86. When the MFEL model trained on samples from the experimental area was applied to other areas, the classification accuracies were 83.94% and 82.77%, respectively. ③ This study verifies the potential application of SAR polarization features and time-sequence harmonic spectral features in mangrove species classification, significantly improving the accuracy for five mangrove species, with accuracies ranging from 76% to 91%. The study results provide valuable insights for expanding the use of medium-resolution remote sensing satellite imagery in monitoring mangrove species.

[Objectives]Currently, systematic research in content retrieval for We-maps is lacking. To address this gap, this paper proposes an approach for geographic feature extraction and retrieval in hand-drawn map scenes using the YOLOv8l-FMSC-Spatial model (You Only Look Once v8l - Fewer Multi-Scale Convolution-Spatial). [Methods]First, different YOLO models were compared to select the optimal YOLOv8l model. The C2f-FMSC module was introduced to improve this model, resulting in the YOLOv8l-FMSC training model specifically designed for We-maps. This model was applied to extract geographic features from raster maps. Next, to meet the retrieval needs of geographic features, a spatial relationship database for these features was established. A spatial computation and retrieval module, Spatial, was designed to process geographic feature information by transmitting and filtering it. The module further calculates spatial correlations between user queries and the geographic feature information in the database. Based on the degree of spatial relationship association, the model indexes maps containing relevant geographic feature information from the We-maps database, enabling the construction of a spatial relationship-based geographic feature retrieval model. The method was validated using hand-drawn campus map retrieval scenarios. The experimental dataset comprised publicly available maps from schools and maps freely created by students, totaling 493 hand-drawn campus maps. These maps were used to study the retrieval of representative geographical elements such as water bodies, sports fields, and unique architectural structures associated with schools nationwide. The focus was on accurately identifying and retrieving these characteristic elements to ensure the model’s practical applicability. [Results] The experimental results indicate: (1) The trained YOLOv8l model effectively identifies geographical elements in self-made maps, with its effectiveness and robustness verified on the proposed dataset; (2)The YOLOv8l model, enhanced with the FMSC module, achieved a precision of 0.8 and a recall of 0.764, making it the optimal choice for practical comparisons; (3)The Spatial calculation model effectively captures the spatial information of relevant geographical elements, narrowing the gap with orthographic map retrieval. By applying this method, the retrieval of geographical elements from hand-drawn campus maps, while considering spatial relationships, becomes achievable. [Conclusions] The proposed model can quickly and accurately retrieve content-relevant hand-drawn maps based on geographic feature conditions, effectively filling the research gap in content retrieval for We-maps.

The accurate extraction of tunnel sections is a pivotal step in tunnel deformation analysis. However, due to inadequate illumination, the reflection and occlusion caused by dust and structural elements generate noise and erroneous points in the point cloud data, complicating data processing. Additionally, the intricate geometry of tunnel features, such as curved surfaces, corners, and cracks, renders traditional section extraction algorithms ineffective for point cloud data. Consequently, there is an urgent need for more efficient and robust algorithms. To address this issue, this paper proposes a method for continuous tunnel section extraction based on laser point cloud data. First, a combination filtering method is introduced, integrating Random Sample Consensus (RANSAC) cylindrical fitting and radius filtering to effectively remove scattered outliers and noise points adhering to the tunnel walls with sparse density. Next, the tunnel central axis is obtained via bidirectional projection, and a mathematical function model is established in line with the principle that ‘a straight line intersecting the tunnel central axis intersects the tunnel wall’, enabling the continuous extraction of tunnel section point clouds. Finally, the fitting center coordinates of the section points and the tunnel design radius are used as parameters to calculate the radial deviation of the tunnel points, representing the shape variables. The tunnel point cloud is visually rendered using these shape variables to display the overall deformation of the tunnel. In this paper, laser point cloud data from three sections of a subway tunnel in Chengdu are used as the experimental data. The results show that the mean values of Class I error, Class II error, and total error are 1.48%, 1.03%, and 1.21%, respectively, with the Kappa coefficient reaching 97.45% when using this method for noise filtering. Compared to traditional methods such as least squares, density clustering, and normal deviation algorithms, this method reduces cumulative errors by 9.34%, 10.61%, and 4.41%, respectively, while increasing the Kappa coefficient by 5.36%, 6.38%, and 3.65%. This demonstrates the enhanced robustness and accuracy of the proposed method. Moreover, the mean deviation between the tunnel section fitting radius obtained through this method and the design radius is merely 1.36 mm, compared to deviations of 1.60 mm and 6.00 mm with existing methods, achieving reductions of 2.5 mm and 2.7 mm, respectively. The range of the tunnel shape variable is between 0 and 18 mm, and the overall deformation of the tunnel is visually displayed through point cloud rendering. The method provides a reliable foundation and essential support for tunnel safety monitoring.

[Objectives] With the development of deep learning technology, the ability to monitor changes in natural resource elements using remote sensing images has significantly improved. While deep learning change detection models excel at extracting low-level semantic information from remote sensing images, they face challenges in distinguishing land-use type changes from non-land-use type changes, such as crop rotation, natural fluctuations in water levels, and forest degradation. To ensure a high recall rate in change detection, these models often generate a large number of false positive change polygons, requiring substantial manual effort to eliminate these false alarms. [Methods] To address this issue, this paper proposes a natural resource element change polygon purification algorithm driven by remote sensing spatiotemporal knowledge graph. The algorithm aims to minimize the false positive rate while maintaining a high recall rate, thereby improving the efficiency of natural resource element change monitoring. To support the intelligent construction and effective reasoning of the spatiotemporal knowledge graph, this study designed a remote sensing spatiotemporal knowledge graph ontology model taking into account spatiotemporal characteristics and developed a GraphGIS toolkit that integrates graph database storage and computation. This paper also introduces a vector knowledge extraction method based on the native spatial analysis of the GraphGIS graph database, a remote sensing image knowledge extraction method based on efficient fine-tuning of the SkySense visual large model, and a polygon purification knowledge extraction method based on the SeqGPT large language model. Under the constraints of the spatiotemporal ontology model, vector, image, and text knowledge converge to form a remote sensing spatiotemporal knowledge graph. Inspired by the manual operation methods for change polygon purification, this paper developed an automatic purification method of change polygons based on first-order logical reasoning within the knowledge graph. To improve the concurrent processing and human-computer interaction, this paper developed a remote sensing spatiotemporal knowledge graph management and service system. [Results] For the task of purifying natural resource element change polygons in Guangdong Province from March to June 2024, the proposed method achieved a true-preserved rate of 95.37% and a false-removed rate of 21.82%. [Conclusions] The intelligent purification algorithm and system for natural resource element change polygons proposed in this study effectively reduce false positives while preserving real change polygons. This approach significantly enhances the efficiency of natural resource element change monitoring.

Addressing the issue of Ultra-Wideband (UWB) signal obstruction by obstacles in indoor environments, which leads to Non Line of Sight (NLOS) errors, this paper presents a fusion positioning method based on Light Detection And Ranging (LiDAR) point cloud for identifying UWB NLOS. This method utilizes LiDAR point cloud information to assist in the identification of UWB NLOS and leverages UWB Line of Sight (LOS) ranging values to eliminate cumulative errors in the LiDAR Simultaneous Localization and Mapping (SLAM) positioning process, thereby enhancing the accuracy and robustness of indoor fusion positioning. Initially, the method processes the LiDAR point cloud using an octree, constructs the ranging direction based on the location information of UWB base stations, and extracts the relevant point cloud data in the ranging direction from the LiDAR point cloud. Subsequently, the 3D Alpha Shape algorithm is employed to extract contours of obstacles that may hinder UWB signal propagation within the extracted point clouds. Furthermore, by analyzing the spatial relationship between the extracted obstacle contours and the UWB ranging direction, the presence of NLOS conditions in UWB signals is effectively determined. Finally, NLOS ranging values identified during the UWB ranging process are excluded, and a tight integration approach is used with an Extended Kalman Filter (EKF) to fuse UWB LOS ranging values with LiDAR SLAM positioning data, eliminating cumulative errors in the LiDAR SLAM positioning outcomes, thereby enhancing the precision and robustness of fusion positioning. Experimental results demonstrate that this method significantly improves positioning accuracy in indoor environments, increasing the positioning accuracy by 96.13% compared to the positioning method that tightly combines the original UWB ranging values with LiDAR SLAM using EKF, with a positioning error of 0.067 m, achieving sub-meter level indoor positioning accuracy.

[Objectives] All meaningful forms of human discourse are rhetorical, and the purpose of rhetoric is to enable communication and foster sympathy between parties with certain views. Narrative maps are essentially a discursive practice for communicating information and exchanging ideas, characterized by the strategic use of rhetoric to construct persuasive discourse and achieve the goal of "agreement" or "persuasion". In the current era, where visual dominance is increasingly prominent, rhetoric has garnered growing attention in cartography. This turn not only addresses core issues in narrative map research but also provides a realistic path for enriching and reconstructing the existing knowledge of modern cartography. However, the academic community has yet to establish a systematic framework, leaving three key issues unresolved: (1) How to conceptualize the rhetoric of narrative maps? (2) How to categorize the rhetoric of narrative maps? (3) What is the working mechanism of rhetoric in narrative maps? [Methods] To address these research gaps, this article, firstly, follows the research paradigm of rhetoric to clarify the essence of rhetoric in narrative maps, and defines it as: "During the design process of narrative maps, cartographers use certain visualization strategies to facilitate the representation of events, thereby weaving explicit narrative intentions into the mapping space in an implicit way to create persuasive discourse or emotional agreement for viewers." Secondly, a classification criterion is proposed based on the differences between content semantic representation and logical semantic representation. Two major categories, semantic rhetoric and structural rhetoric, along with 24 minor classes, are divided for rhetoric of narrative map. Semantic rhetoric mainly focuses on enhancing the understanding of content, expressing the connotation and imaginative tension of map "text". Structural rhetoric aims to emphasize the logic semantic relationships in narrative discourse, presenting the narrative logic of events. Semantic rhetoric often manifests as the design of visual symbols to describe events, serving as the "visual punctum" of narrative maps. Structural rhetoric typically involves adjusting the arrangement and structure of different event units, functioning as the "visual stadium" of narrative maps. Next, the mechanism of rhetoric in narrative maps is explored from four aspects: the dimensions of rhetoric, the hierarchy of rhetoric, the integrated use of rhetoric, and the applicability principles of rhetoric. Finally, this study demonstrates the applicability of the proposed theoretical framework through a case study of "Jiangnan Canal", illustrating how the framework can facilitate narrative map design. [Conclusions] This paper lays a theoretical foundation for narrative map research and contributes to the theoretical innovation of contemporary cartography.

[Objectives] Remote Sensing Intelligent Interpretation (RSII) often encounters challenges when applied for practical resource and environmental management, especially for complex scenes. To address this, we start from the explanation of why remote sensing interpretation is needed, and clarify that the mission of RSII is to achieve more rapid interpretation to build the digital twin earth with lower cost compared to manual interpretation. However, most RSII systems operate as a unidirectional process from remote sensing data to geoscience knowledge, lacking the feedback from knowledge to data. As a result, remote sensing information extracted from data often mismatch the knowledge of existing geoscience, creating a trust crisis between RSII researchers and geoscience researchers. And the crisis becomes more severe with the uncertainty of remote sensing information. [Analysis] We believe that an agreed upon representation model of geoscience knowledge between RSII researchers and geoscience researchers is necessary to alleviate the crisis. Based on this analysis, we propose a framework using geo-science zoning as the bridge to connect RSII researchers and geoscience researchers. In this framework, knowledge from geoscience could be transferred into the RSII system through geo-science zoning so that the interpretation results could be more coincided with geoscience knowledge. The framework mainly relies on (a) the scene complexity measurement, (b) the knowledge coupling of geographic regions to form the geological zoning method for remote sensing intelligent interpretation, and (c) the sampling specification of regional samples. The scene complexity measurement provides quantitative features for geoscience zoning and sampling weights assignment. Existing zoning data, such as ecological zoning data, geographic elements, and multisource remote sensing images are the main data inputs for geoscience zoning. The main principles for constructing zoning methods include (a) the geoscience elements type, (b) the scale of geoscience zoning, and (c) the process of information flow from data to knowledge. [Prospects] With these models, we can realize regional RSII guided by the knowledge. Preliminary experiments on complexity and optimization sampling, image segmentation scale optimization, cultivated land type fine classification, etc., reveal that this framework has great potential in improving the geoscience knowledge acquisition by RSII, enhancing the accuracy of the state-of-the-art RSII by 6%~10%, especially for the high-complexity nature scenes. However, the superiority of the framework may disappear if the scene for interpretation is simple, like the first level land use/cover classification, which is mainly caused by the inefficient samples after geoscience zoning. Therefore, more attention is needed in sampling when developing geoscience zoning framework.

Bicycle sharing offers the advantages of resource sharing, environmental sustainability, and low carbon emissions, and has been widely applied in recent years. Trajectory prediction of shared bicycles is crucial for the scientific and efficient planning of infrastructure. However, existing trajectory prediction mechanisms are relatively limited, and the influencing factors are often too narrow, leading to low prediction accuracy. This restricts the further growth and development of bicycle-sharing systems. Therefore, accurately predicting the travel trajectories of shared bicycles is essential for optimizing bicycle lanes, efficiently deploying and scheduling bicycle resources, improving road design, and addressing the "last mile" challenge in urban transportation. To better understand the spatio-temporal characteristics of shared bicycle travel and the influence of natural and weather factors on travel trajectories, and to improve prediction accuracy, this paper developes a trajectory prediction model that integrates natural and weather factors with a spatio-temporal attention residual bi-directional network (NWSTAR-BiLSTM). This study uses shared bicycle order and trajectory data from Xiamen, provided by the government’s open data platform, to analyze the spatio-temporal distribution of travel and the impact of natural and weather factors on trajectories. The model incorporates variables such as temperature, weather conditions, wind speed, and air quality, dividing the shared bicycle trajectory data into time series based on periodicity. Using an attention mechanism and residual learning, the prediction results are adjusted according to weather factors. The dataset is divided into training, testing, and validation sets in a 7:2:1 ratio, and the model undergoes training, parameter adjustment, and comparative validation. Experimental results show that the trajectory prediction accuracy of the NSTARWSTAR-BiLSTM model exceeds that of traditional models, such as LSTM, BiLSTM, CNN, Att-LSTM, and self-built comparative models (e.g., STAR-BiLSTM without natural and weather factors, WSTR-BiLSTM without the attention mechanism, and WSTA-BiLSTM without the residual network). The NSTARWSTAR-BiLSTM model not only inherits the strengths of traditional residual network models but also innovatively integrates the attention mechanism with multiple natural and weather factors, enhancing trajectory prediction accuracy. The model also demonstrates strong intelligent learning capabilities, with prediction accuracy improving as feedback increases.

[Objectives] Ship detection using Synthetic Aperture Radar (SAR) images has gained widespread recognition and application across various fields, including marine search and rescue, port reconnaissance, and territorial sea defense. Nevertheless, with the rapid advancement of on-orbit intelligent processing technologies, higher demands have emerged for real-time detection of ship targets in spaceborne SAR images. [Methods] To address challenges such as the diverse scales of ship targets in current SAR images, the complex background of shore-based vessels, and the limited hardware resources of various remote sensing platforms, this paper presents a lightweight SAR image ship detection model, LWM-YOLO. Firstly, we propose a Lightweight Backbone Network (LWCA) designed specifically for SAR image processing. The LWCA integrates an optimized backbone network with an attention mechanism, effectively reducing the model's complexity and parameter size while maintaining high performance and lowering computational demands. Secondly, to tackle the issue of diverse target scales in SAR images, we have constructed a lightweight feature fusion module, termed LGS-FPN. This module enhances the extraction of detailed information on ship targets in SAR images by efficiently fusing features from different scales, improving detection performance for ship targets of various sizes. Furthermore, the module minimizes computational complexity, ensuring that the model can operate smoothly without significant resource consumption. In addition to addressing the scale issue, we have also focused on optimizing localization accuracy. We introduce a detection architecture based on the MPD-Head, which leverages the strengths of the MPD-Head to improve detection performance for small ship targets in complex environments. Finally, we validate the proposed algorithm through comparative experiments with mainstream methods on the LS-SSDD and SSDD ship detection datasets. [Results] The results demonstrate that our algorithm achieved mean Average Precision (mAP) values of 74.7% and 97.3% on the respective datasets, representing improvements of 1.5 and 1.0 percentage points over the baseline model. Additionally, the parameter size of our model was reduced to 36% of the baseline model, and computational complexity decreased to 80%. [Conclusions] Compared to other mainstream algorithms, the proposed method demonstrates not only higher accuracy but also significant advantages in detection speed. These findings can provide robust support for intelligent target detection, space-based in-orbit applications, and related fields.

[Objectives] City-wide traffic flow prediction plays a crucial role in intelligent transportation systems. Traditional studies partition road networks into grids, represent them as graph structures with grids as nodes, and use graph neural networks for region-level prediction. However, this region-based approach overlooks the relationships between individual roads, making it difficult to reflect traffic flow changes of roads. Methods based on road segment data can better capture spatial connections between roads and enable more accurate traffic flow predictions. However, mapping trajectory data to roads presents challenges such as redundant data and trajectory mismatches, and traffic flow data after mapping is sparse. Existing methods struggle to effectively capture the spatial correlation in sparse traffic conditions. [Methods] To address these issues, this study proposes an Attention Spatio-Temporal Neural Network (ASTNN) model for road-level sparse traffic flow prediction. The model first preprocesses trajectory data and applies Hidden Markov Model (HMM)-based map matching to obtain road-level traffic flow data. It then introduces an adaptive compact 2D image representation method to model the road network as a 2D image, where road segments are represented as pixel points. Based on an analysis of the spatial and temporal characteristics of traffic flow, two new attentional spatiotemporal blocks are proposed: Attentional Spatio-Temporal Memory Block (ASTM block) for mining temporal correlations and attentional spatial-temporal focusing block (ASTF block) for extracting spatial sparse features. By integrating these two blocks with external information, ASTNN is constructed to achieve road-level traffic flow prediction. [Results] This study uses Chengdu taxi trajectory data as a case study. After preprocessing trajectory data and mapping traffic flow, the proposed model is validated on a five-level road network within Chengdu’s third ring area. Results indicate that the proposed data processing method reduces trajectory-to-road network matching time by 73.6%. In the comparative experiments with existing models, such as Convolutional Neural Network (CNN), Convolutional Long Short-Term Memory (ConvLSTM), Gated Recurrent Unit (GRU), and Spatial-Temporal Neural Network (STNN), ASTNN achieves the highest prediction accuracy in terms of Root Mean Square Error (RMSE), Mean Absolute Error (MAE), and R-squared (R²). Furthermore, the study confirms the significant improvement in prediction accuracy when incorporating temperature data into ASTNN, providing new insights for optimizing model performance. [Conclusions] The ASTNN model proposed in this study provides an effective framework for city-wide, road-level sparse traffic flow prediction, offering valuable insights for intelligent transportation systems.

[Objectives] With accelerating urbanization and a surge in vehicle numbers, urban traffic systems face immense pressure. Intelligent transportation systems, a vital component of smart cities, are widely employed to improve urban traffic conditions, with traffic speed prediction being a key research focus. However, the complex coupling relationships and dynamically varying characteristics of urban traffic network nodes pose challenges for existing traffic speed prediction methods in accurately capturing dynamic spatio-temporal correlations. Spatio-temporal graph neural networks have proven to be among the most effective models for traffic speed prediction tasks. However, most methods heavily rely on prior knowledge, limiting the flexibility of spatial feature extraction and hindering the dynamic representation of road network topology. Recent approaches, such as adaptive adjacency matrix construction, address the limitations of static graphs. However, they often overlook the synergy between dynamic features and static topology, making it difficult to fully capture the complex fluctuations in traffic flow, which in turn limits prediction accuracy and adaptability. [Methods] To address these challenges, this study formulates urban traffic speed prediction as a multivariate time-series forecasting problem and proposes a traffic speed prediction model based on a Multivariate Time-series Dynamic Graph Neural Network (MTDGNN). Leveraging real-time traffic information and predefined static graph structures, the model adaptively generates dynamic traffic graphs to capture spatial dependencies through a graph learning layer and integrates them with static road network graphs to capture spatial dependencies from multiple perspectives. Meanwhile, the alternating use of graph convolution and temporal convolution modules constructs a multi-level spatial neighborhood and temporal receptive field, fully exploring the spatial and temporal features of traffic data. [Results] The MTDGNN model was tested on real traffic data from 397 road sections in eastern Beijing, collected between April 1, 2017, and May 31, 2017. Its prediction results were compared against nine benchmark models and seven ablation models. Compared to benchmark models, MTDGNN reduced the average MAE by at least 2.24% and the average RMSE by at least 3.98%. [Conclusions] Experimental results demonstrate that the MTDGNN model achieves superior prediction accuracy in MAE, RMSE, and MAPE evaluation metrics, highlighting its robustness and effectiveness in complex traffic scenarios.

[Objectives] Cancer is the leading cause of death in most countries worldwide, posing a significant threat to human longevity and public health. This study explores the spatiotemporal distribution characteristics of mortality rates for five major types of cancer worldwide and provides predictions for future trends. [Methods] Aiming at five major cancer types (lung, colorectal, gastric, liver, and pancreatic cancer) in 200 countries from 2011 to 2019, this study used GBD and World Bank data to extract spatial heterogeneity of the factors affecting cancer mortality using the MGWR model. The ARIMA model was used to extract temporal trend characteristics of various cancer mortality rates. Such spatial-temporal information was integrated into a Bayesian spatial-temporal model to predict and evaluate the global mortality risk for the five types of cancer. [Results] Results revealed that global death rate for all five cancer types increased, with an average rise of 17.2 deaths per 100 000 people in 2019 compared to 2011. Over 72.8% of countries exhibited a high relative risk of cancer death (RR>1), indicating significant spatial clustering. [Conclusions] Regions such as Europe, Central Asia, North America, and East Asia and the Pacific experienced faster increases in cancer death rates compared to Africa and South Asia. Compared to low- and middle-income countries, middle-high- and high-income countries showed a more pronounced upward trend in cancer mortality and a higher relative risk. Key factors influencing global cancer mortality included the percentage of the population aged 65 years and older, smoking, alcohol consumption, low physical activity, high sugar diets, GDP per capita, GNI per capita, and health expenditure per capita. By integrating the advantages of different geographical spatial-temporal analysis methods, this study developed an innovative spatiotemporal prediction model of disease risk that integrates spatial-temporal grouping variables and multiple influencing factors. This proposed model is highly flexible, interpretable, and better suited for quantifying non-stationarity spatial-temporal relationships. While the structured spatial and temporal effects increase computational demands, the model effectively assesses cancer mortality risk across regions, offering robust insights into the spatiotemporal dynamics of disease. This approach deepens the integration of geospatial modeling technology and epidemiological research, providing significant scientific contributions to global cancer research, prevention, and control planning.

[Significance] As a comprehensive observation of natural resource development and utilization, spatio-temporal big data on natural resources contains valuable knowledge about resource distribution, spatio-temporal process evolution, and interrelationships. [Progress] This paper examines spatio-temporal big data mining and knowledge services for natural resources, highlighting key data mining techniques and their critical applications in knowledge services. First, it introduces the core concepts, technical frameworks, and methodological processes of spatio-temporal clustering analysis, association mining, anomaly detection, predictive modeling, and geographic risk assessment, along with their applications in natural resource management and land-use decision-making. Second, a four-tier natural resource spatio-temporal knowledge service system is proposed, encompassing descriptive, diagnostic, predictive, and decision-making knowledge services, which provide essential support for applications such as resource status monitoring, land-use regulation, and disaster prevention and mitigation. Finally, the paper indicates that current natural resource management is transitioning from data aggregation and analysis to knowledge-driven intelligent services, forming an emerging research and application paradigm of big data, big analysis, big knowledge, and big services. [Prospect] Future efforts will focus on advancing collaborative data and knowledge mining technologies, addressing the standardization challenges in spatio-temporal knowledge bases and services, and exploring the potential of cutting-edge technologies such as generative large models in the natural resource domain to drive the information and intelligent transformation of natural resource management.

[Significance] The extraction of Cartographic-Level Vector Elements (CLVE) is a critical prerequisite for the direct application of remote sensing image intelligent interpretation in real-world scenarios. [Analysis] In recent years, the continuous rapid advancement of remote sensing observation technology has provided a rich data foundation for fields such as natural resource surveying, monitoring, and public surveying and mapping data production. However, due to the limitations of intelligent interpretation algorithms, obtaining the necessary vector elements data for operational scenarios still heavily relies on manual visual interpretation and human-computer interactive post-processing. Although significant progress has been made in remote sensing image interpretation using deep learning techniques, producing vector data that are directly usable in operational scenarios remains a major challenge. [Progress] This paper, based on the actual data needs of operational scenarios such as public surveying and mapping data production, conducts an in-depth analysis of the rule constraints for different vector elements in remote sensing image interpretation across a wide range of operational contexts. It preliminarily defines "cartographic-level vector elements" as vector element data that complies with certain cartographic standard constraints at a specific scale. Centered on this definition, the content of the rule set for CLVE is summarized and analyzed from nine dimensions, including vector types, object shapes, boundary positioning, area, length, width, angle size, topological constraints, and adjacency constraints. Evaluation methods for CLVE are then outlined in four aspects: class attributes, positional accuracy, topological accuracy, and rationality of generalization and compromise. Subsequently, through literature collection and statistical analysis, it was observed that research on deep learning-based vector extraction, while still in its early stages, has shown a rapid upward trend year by year, indicating increasing attention in the field. The paper then systematically reviews three major methodological frameworks for deep learning-based vector extraction: semantic segmentation & post-processing, iterative methods, and parallel methods. A detailed analysis is provided on their basic principles, characteristics and accuracy of vector extraction, flexibility, and computational efficiency, highlighting their respective strengths, weaknesses, and differences. The paper also summarizes the current limitations of remote sensing intelligent interpretation methods aimed at CLVE in terms of cartographic-level interpretation capabilities, rule coupling, and remote sensing interpretability. [Prospect]Finally, future research directions for intelligent interpretation of CLVE are explored from several perspectives, including the construction of broad and open cartographic-level rule sets, the development and sharing of CLVE datasets, the advancement of multi-element CLVE extraction frameworks, and the exploration of the potential of multimodal coupled semantic rules.

[Objectives] Cross-view image matching and localization refers to the technique of determining the geographic location of a ground-view query image by matching it with a geotagged aerial reference image. However, significant differences in geometric appearance and spatial layout between different viewpoints often hinder traditional image matching algorithms. Existing methods for cross-view image matching and localization typically rely on Convolutional Neural Networks (CNNs) with fixed receptive fields or Transformers with global modeling capabilities for feature extraction. However, these approaches fail to fully address the scale differences among various features in the image. Additionally, due to their large number of network parameters and high computational complexity, these methods face significant challenges in lightweight deployment. [Methods] To address these issues, this paper proposes a lightweight cross-view image matching and localization method that employs multi-scale feature aggregation for ground panoramic and satellite images. The method first extracts image features using LskNet, then designs and introduces a multi-scale feature aggregation module to combine image features into a global descriptor. The module decomposes a single large convolution kernel into two sequential smaller depth-wise convolutions, enabling multiple scale feature aggregation. Meanwhile, spatial layout information is encoded into the global feature, producing a more discriminative global descriptor. By integrating LskNet and the multi-scale feature aggregation module, the proposed method significantly reduces parameters and computational cost while achieving superior accuracy on publicly available datasets. [Results] Experimental results on the CVUSA, CVACT, and VIGOR datasets demonstrate that the proposed method achieves Top-1 recall rates of 79.00% and 91.43% on the VIGOR and CVACT datasets, respectively, surpassing the current highest-accuracy method, Sample4Geo, by 1.14% and 0.62%. On the CVUSA dataset, the Top-1 recall rate reaches 98.64%, comparable to Sample4Geo, but with parameters and computational costs reduced to 30.09 M and 16.05 GFLOPs, representing only 34.36% and 23.70% of Sample4Geo's values, respectively. Additionally, ablation experiments on public datasets show that the multi-scale feature aggregation module improves the Top-1 recall rate of the baseline network by 1.60% on the CVUSA dataset and by 13.48% on the VIGOR dataset, further validating the effectiveness of the proposed method. [Conclusions] Compared to existing methods, the proposed algorithm significantly reduces both parameters and computational costs while maintaining high accuracy, thereby lowering hardware requirements for model deployment.