ABSTRACT Online geocoding platforms transform textual addresses into geographic coordinates, yet often exhibit substantial geocoding errors. Most multi‐source geocoding approaches lack explicit spatial constraints, limiting the stability and interpretability of fused results. To address this issue, this study proposes a Polygon Constraint‐based Multi‐source Geocoding optimization method (PCMG) that integrates geocoding outputs from multiple platforms under explicit polygon constraints corresponding to address semantics. The method aggregates candidate coordinates and applies spatial filtering and POI‐based refinement to enhance spatial plausibility and consistency. It is evaluated using 1884 addresses in Nanjing, China, with ground‐truth locations collected through field surveys. Results demonstrate that PCMG significantly reduces mean error and suppresses extreme positional deviations. Error distributions are more concentrated and exhibit a clearer relationship with polygon scale, indicating improved spatial consistency. The findings suggest that polygon units can serve as active constraints for multi‐source geocoding optimization and improve the reliability of online geocoding services in complex urban environments.
ABSTRACT A GIS‐based analysis of invariant spectral‐activity trajectories in a long‐term backbone transmission network. Understanding the long‐term evolution of geographic networks requires methods that separate structural change from mechanical changes in network size and connectivity. We analyze annual snapshots of the Russian backbone power transmission network from 1933 to 2020 using component‐wise normalized‐Laplacian spectral activity, temporal trajectory descriptors, geodesic spatial statistics, and GIS‐based areal modeling. A Gaussian mixture model identifies eight substantial trajectory types among 1613 nodes (92–372 nodes per type; mean bootstrap adjusted Rand index 0.753). Global Moran's I demonstrates positive spatial autocorrelation for mean activity (0.137), mean annual change (0.130), maximum annual change (0.077), and the weaker long‐term trend (0.035). An eight‐class join‐count test finds 3304 same‐type neighbor joins versus 1824 expected under random labeling ( p = 0.0001). An uncertainty‐aware regionalization framework combines class‐conditional adaptive Gaussian kernels, equal class priors, geodesic support limits, and cell‐level probability diagnostics to translate the point typology into 10 km model‐based zones. Spectral activity is most strongly associated with inverse‐length node strength (Spearman ρ = 0.748), but only weakly with betweenness ( ρ = 0.223), confirming that it captures a distinct dimension of structural embedding.
ABSTRACT Buildings are the primary disaster‐bearing structures of natural disasters such as earthquakes, and accurately diagnosing the 3D geometric morphology of damaged buildings is crucial for emergency response and post‐disaster reconstruction. However, existing methods rely solely on pure point cloud geometric alignment and do not incorporate multidimensional domain knowledge, such as disaster, construction, and diagnostic standards, resulting in significant shortcomings in accuracy and efficiency when applied to complex damage scenes. To address these challenges, we propose a knowledge‐guided and feature‐constrained method for accurate 3D geometric morphology diagnosis of damaged buildings. First, we analyze the characteristics and interrelationships of damaged building scenes, constructing a multi‐domain associated scene knowledge graph (KG) based on “event‐object‐morphology‐rule”. Second, geometric features of buildings are extracted from multi‐temporal point clouds, and a precise comparison algorithm for 3D geometric morphology of damaged buildings constrained by linear and surface features is designed. Then, a scene knowledge‐guided building damage assessment method is proposed, and building geometric features along with associated rules are analyzed. Finally, a damaged building is selected as a case area for experimental analysis. Experiments show that the proposed method maintains a diagnostic accuracy above 98% for 3D geometric morphology, which is at least 10% higher than conventional methods, achieving precise diagnosis of 3D geometric morphological changes in damaged buildings. Furthermore, compared to the ICP and GeoTransformer, our approach improves diagnostic efficiency by over 80%, enabling timely and effective information support for post‐disaster emergency response and reconstruction, while also establishing a new paradigm for 3D geometric morphology diagnosis of damaged buildings.
ABSTRACT The siting of neighborhood‐scale service facilities requires balancing accessibility, facility capacity, and feasible land under environmental and regulatory constraints. Municipal solid waste collection points provide a representative case, especially across contrasting urban and rural settlement forms. Existing approaches often address only parts of this problem: GIS‐based suitability analysis identifies candidate areas but rarely incorporates explicit assignment and capacity control, while clustering and location‐allocation methods may overlook siting feasibility under exclusion constraints. This paper develops an integrated geospatial siting framework, the Spatial Capacitated Allocation Network (SCAN), which combines capacitated assignment, soft service radius, and exclusion‐aware location refinement within a single workflow. SCAN is solved through an Alternating Assignment Relocation procedure integrating mixed‐integer assignment with continuous relocation while preserving siting feasibility. Two case studies from Shenzhen and Sichuan show that SCAN generates feasible locations and balanced allocations while revealing how settlement structure shapes accessibility and feasibility trade‐offs.
ABSTRACT Daily operations in large campuses depend on how efficiently people move through space and time. In this sense, course timetables are more than administrative schedules: they act as mobility policies that orchestrate thousands of trajectories, shaping travel burden, congestion, accessibility, and the reliability of back‐to‐back transitions. Designing timetables that are both feasible and mobility‐friendly is challenging because hard constraints such as capacity, conflicts, and feasibility must be satisfied alongside soft constraints such as preferences, satisfaction, and coordination, all under dynamic conditions such as real‐time disruptions and evolving demand. Traditional static optimization methods often struggle to capture these human mobility impacts and to adapt when campus conditions change. This paper reconceptualizes course timetabling as a recommendation‐based task and leverages the Texas A&M Campus Digital Twin as a dynamic data platform to evaluate mobility consequences at scale. We propose an iterative framework that combines data‐driven candidate generation, multi‐criteria mobility‐aware scoring, and feedback‐driven refinement to generate ranked sets of adaptive timetable recommendations. A mobility‐aware composite scoring function combining classroom occupancy, travel distance, travel time, and vertical transitions systematically balances resource efficiency with human‐centered movement costs. Extensive experiments using real‐world data from Texas A&M University show that the proposed approach reduces mobility friction and travel inefficiencies and improves classroom utilization. By coupling recommendation‐oriented decision‐making with digital twin intelligence, this study provides a robust and scalable blueprint for mobility‐centered campus planning and resource allocation, with potential extensions to broader urban systems.
ABSTRACT Efficient and precise toponym and address retrieval has become a cornerstone for sectors such as logistics, intelligent navigation, and urban planning. However, conventional address retrieval methods often struggle with semantic ambiguity, incomplete information, or nonstandard address formats, resulting in suboptimal retrieval accuracy and limited robustness. To tackle these challenges, this study proposes AddressRAG, an innovative intelligent retrieval framework that integrates Knowledge Graphs (KG) with Large Language Models (LLM). AddressRAG employs a custom Semantic Text Augmentation (STA) mechanism to enhance the administrative hierarchy and spatial constraints of the raw data. Then we construct an enhanced hierarchical knowledge graph based on prompt tuning. Furthermore, a hierarchical community reporting mechanism was introduced for AddressRAG to accurately capture macro‐level administrative constraints and micro‐level spatial relationships. Experiments on a cross‐city dataset showed that AddressRAG obtained higher observed mean scores in faithfulness, answer relevance, candidate retrieval, and administrative hierarchy consistency, with uncertainty quantified using bootstrap confidence intervals. This research provides critical technical support for scalable and adaptive toponym address retrieval.
ABSTRACT With the development of positioning technologies and Location‐based service (LBS) applications, trajectory big data has been rapidly accumulated. It has become a key resource for understanding urban dynamics, traffic patterns, and mobility behaviors. Trajectory similarity measurement serves as a fundamental task underpinning various applications, such as user identification, trajectory clustering, and anomaly detection. However, existing similarity measurement methods face significant challenges in handling the inherent spatiotemporal heterogeneity and multi‐granularity of trajectory data. It remains difficult for these methods to capture diverse mobility behavioral patterns within complex trajectories and to coordinate the processing of point‐level details and sequence‐level semantic information. To address this issue, this paper proposes a Mixture‐of‐Experts Enhanced Transformer for Trajectory Similarity Measurement (MoE‐TrajSim). Specifically, MoE‐TrajSim adopts a contrastive learning‐based encoder‐decoder framework. The encoder incorporates a Mixture‐of‐Experts (MoE) layer instead of a standard Feed‐Forward Network (FFN), and MoE uses five specialized experts to dynamically capture heterogeneous trajectory patterns. By introducing a point‐aware sequence‐level routing strategy, the MoE router facilitates the seamless integration and synergy between point‐level features and sequence‐level semantic information. Experimental results on three real‐world datasets demonstrate that MoE‐TrajSim outperforms state‐of‐the‐art baselines in various metrics, effectively improving accuracy and robustness under noise. Ablation experiments, grid‐resolution sensitivity analysis, and cross‐dataset transfer evaluation further verify the effectiveness, structural rationality, spatial discretization robustness, and generalization capability of MoE‐TrajSim. The proposed MoE‐TrajSim may provide a new solution for measuring the similarity of complex, heterogeneous trajectories and also hold important implications for advancing the practical applications of trajectory data mining.
ABSTRACT Public service facilities (PSF) are fundamental indicators of urban livability and sustainable development. Their spatial configuration critically influences operational efficiency and residents' quality of life. Nevertheless, achieving an optimal distribution and elevated service level within limited urban space remains a pressing challenge. In this study, we leverage point of interest data (2017–2022) across 374 townships in Chengdu to evaluate PSF performance through a tripartite framework. We analyze six essential service categories: healthcare (HF), education (EF), transportation (TF), commercial services (CSF), parks and recreation (PRF), and food services (FSF) to uncover spatiotemporal dynamics, spatial differentiation, and clustering behavior. Our findings reveal substantial overall improvement in PSF provision over the study period, though significant spatial inequities persist. A pronounced “core‐periphery” gradient was identified, with central townships exhibiting PSF 3.5 times higher than those in peripheral regions. Central areas consistently led in adequacy, while equity improvements were observed across all facility types. TF was the most equitable, followed by EF and PRF. Convenience remained moderate across most townships, with persistently low scores in mountainous zones such as Longmen and Longquan. The evolutionary trajectory of PSF clustering manifested four distinct phases: fragmentation, polarization, diffusion, and maturation. Based on these patterns, we propose spatially targeted planning strategies to enhance service distribution and reinforce urban–rural linkages. This study provides an empirical basis for evaluating Chengdu's PSF policies and introduces a multidimensional framework for microscale spatial analysis, thereby offering valuable support for the promotion of public service equity in Chengdu and other megacities.
ABSTRACT Birbhum, one of the under‐served districts of Eastern India, suffers immensely from limited healthcare accessibility and inadequate emergency obstetric care service coverage. The study aims to identify suitable sites for new hospitals using a hybrid geospatial and Multi‐Criteria Decision‐Making (MCDM) framework to address existing service coverage gaps. An integrated framework combining Analytical Hierarchy Process (AHP), Multi‐Attribute Utility Theory (MAUT), and Entropy‐CRITIC based weighting techniques was employed. The findings indicated that the Entropy–MAUT approach demonstrated comparatively better spatial correspondence with existing hospital locations, as 24% of them fell within the highest suitable class, while a comparatively lower proportion (19%) fell within the low suitability class. In contrast, both AHP and CRITIC–MAUT approaches identified no hospitals in the highest suitability class. Notably, one‐third (33%) of hospitals were located outside all defined suitability classes across methods, indicating a marked mismatch between prevailing healthcare infrastructure and community needs. The results highlighed the critical role of accessibility and connectivity in optimal hospital site selection, as these factors determined how easily patients could avail themselves of facilities through the surrounding transport network.
ABSTRACT Physiological arousal, as a typical objective perception, can avoid the influence of human subjective biases and directly express emotional intensity. Existing studies on physiological arousal were limited to small‐scale urban areas, lacking the exploration of large‐scale estimation methods and the spatial heterogeneity of physiological arousal. This paper proposes a framework for analyzing physiological arousal by combining street view image (SVI) and electrodermal activity (EDA). An interpretable machine learning method was applied to reveal the impact of visual factors on physiological arousal. By capturing the neighboring visual environment, a hybrid deep learning model was constructed to estimate physiological arousal. Our results revealed that horizontal vegetation and soil exerted a stronger modulatory effect on physiological arousal, while traffic signs reduced arousal, contrasting with arousal enhancement from traffic lights. The model accounting for the neighboring visual environment outperforms the single‐SVI model. These findings provide an “objective” perspective for urban planning and mental health interventions.
ABSTRACT The term walkability refers to the degree to which pedestrians are willing to walk along a street. However, walkability is a broad concept that includes various subcategories and assessment methods. In this article, we present a literature review on a less commonly studied aspect of walkability: “Visual Walkability”. Visual walkability refers to the walkability perceived by pedestrians through visual stimuli in the urban environment. With technological advancements, assessing walkability has become easier, shifting from field research to the increasing use of computer vision, particularly for analyzing visual elements. Here, we applied the PRISMA method, reviewed 63 articles, and summarized and compared the assessment methods, types of information, and explanatory variables used in these studies. We also highlight existing limitations in visual walkability research, such as the lack of validation of results, the limitations of car‐oriented SVI services, and disagreements among researchers regarding the influence of certain variables. Finally, we suggest future directions for research, which may lead to further developments in the field of visual walkability.
ABSTRACT Advances in understanding the interactions between the built environment (BE) and traffic spatiotemporal congestion patterns (TCPs) have significantly contributed to sustainable transportation development. However, existing BE‐TCP studies remain fragmented in terms of measurement, built environment indicators, spatial scale and methodology selection, which restrict cross‐study comparability and policy generalization. To address these gaps, this review systematically analyses 214 papers collected from Web of Science, Scopus, and Google Scholar published between 1997 and 2025. This review makes three main contributions. First, it synthesizes the BE measurement from the conventional 5Ds framework (density, diversity, design, distance to transit, and destination accessibility) to an expanded 7Ds by incorporating two additional dimensions: demand management and demographics. Second, it identifies four hierarchical geospatial scales of BE‐TCP relationships—disaggregated, neighborhood, aggregated, and regional scales, and highlights the scale‐dependent nature of BE‐TCP effects. Third, it summarizes the methodological evolution from classical linear to spatial nonlinear, from direct to mediation, from singular to synergic, from static to dynamic, and from correlation to causality approaches. The review reveals that BE‐TCP relationships are not uniform but are characterized by strong spatiotemporal heterogeneity, multi‐scale effects, and regional dependence. Based on these advances, three critical challenges and four opportunities are proposed. Overall, this review highlights the need to move BE‐TCP research from fragmented empirical associations toward spatiotemporal, multiscale, and context‐sensitive explanations that can better support transferable and locally adaptive policy‐making.
Natural resource governance relies on standardized specifications for systematic management, yet current Natural Resource Standards and Specifications (NRSS) suffer from fragmentation, outdated content, and inadequate integration, hindering their synergistic application. This study proposes a semantic harmonization framework that integrates fragmented standards through a large language model-driven knowledge graph. We develop a multi-granular knowledge structure comprising four hierarchical layers: knowledge source layer, core concept layer, secondary concept layer, and foundation layer. The framework employs few-shot prompting for document-level knowledge extraction and Chain-of-Thought reasoning for clause-level knowledge extraction. Experimental results demonstrate that our method significantly outperforms baseline models, with the Qwen3-8B model achieving performance comparable to or exceeding the Qwen3-32B model on key metrics. The proposed framework effectively consolidates fragmented standard resources, enhances semantic understanding, and provides precise knowledge services for natural resource governance tasks including surveys, monitoring, rights confirmation, asset accounting, and ecological protection. This work offers a technical pathway for the digital transformation and intelligent application of NRSS, supporting high-quality development in resource management.
Urban building clustering is essential for deciphering spatial structures and functional patterns. Yet, existing methods mainly rely on local or pairwise geometric similarity measures, which fail to capture the latent nonlinear topological structure of building layouts and insufficiently handle structural inconsistencies across multi-view data, leading to compromised robustness. To address these issues, this study proposes a novel multi-view clustering model that, for the first time, integrates topological manifold modeling and cross-view diversity detection in a unified framework. This model first constructs three complementary feature views (centroid distance, outline distance, and non-spatial attributes) and models latent topological manifolds within each view. A novel diversity detection mechanism is then proposed to identify and suppress inconsistent structural information both within individual views and across different views, producing a pure graph for each view, which is then fused into a consensus graph with a clear clustering structure. Finally, an alternating iterative optimization algorithm is proposed to jointly learn topological correlation, multi-view consistency, and consensus structure within a unified framework. Extensive experiments on 15 urban communities in Wuhan and Chengdu, China, show that our model consistently outperforms 12 state-of-the-art baselines, achieving up to 32% improvement in Adjusted Rand Index (ARI) and over 20% gains in Accuracy and F-score. In particular, linear patterns achieve an average accuracy of over 90%.
ABSTRACT Regression‐based local modeling through multiscale geographically weighted regression (MGWR) is popular in studies of spatial processes across a wide array of disciplines, and six major software packages exist for this purpose: MGWR 2.2.1; mgwr‐plugin 1.1 in QGIS; MGWR in ArcGIS Pro 3.4.3; GWmodel 2.4.1 in R; GWmodelS 1.0.4; and mgwrsar 1.1 in R. We compare major functionalities, model specification options, and performance attributes of these algorithms using both simulated and empirical data and show that the calibration results are not always consistent. The results provide an important reminder that the software used in calibrating local models should always be cited.
With the rapid development of digital technologies and participatory mapping platforms, crowdsourced geographic data has emerged as a valuable and widely adopted resource for transportation research and urban planning. Broadly, crowdsourced geographic data refers to geospatial datasets collected, edited, and shared by the general public via a crowdsourced contribution model, which is a core form of Volunteered Geographic Information (VGI). Crowdsourced highway network data, as a key application subset of crowdsourced geographic data, is the core research object of this study. OpenStreetMap (OSM), as one of the most well-established and widely used crowdsourced geographic data platforms globally, serves as the primary data source for this work. However, the open, voluntary user-generated nature of OSM brings inherent challenges to data quality, with commonly noted issues including missing features, geometric deviations, and inaccurate attribute information. To date, most prior studies on OSM road network data quality have predominantly focused on urban areas, where higher contributor density generally supports more stable data quality. In comparison, the quality of highway and intercity road networks has received relatively limited research attention, despite the more prominent quality-related issues observed in these datasets. Meanwhile, existing related studies in the Chinese context are mostly limited to local or regional spatial scales with short observation periods, and relatively few have established a systematic evaluation framework aligned with international geospatial data quality standards or conducted quantitative analysis of the factors driving spatial and temporal variations in OSM highway data quality. To address these noted gaps in existing research, this study evaluates the quality of OSM highway network data across China, with a spatial scope covering 366 prefecture-level cities nationwide and a continuous temporal window spanning 2015-2024. By constructing matched crowdsourced datasets and authoritative reference datasets, we develop a comprehensive evaluation framework aligned with international geospatial data quality standards, which covers four key dimensions of data quality: completeness, positional accuracy, logical consistency, and temporal validity. This framework complements existing fragmented, single-dimensional evaluation approaches by enabling a systematic multi-faceted assessment of highway network data quality. Based on this framework, we analyze the temporal evolution, spatial clustering patterns, and statistical distribution characteristics of the four quality indicators over the study period. To further explore the drivers of spatial and temporal variations in data quality, we compile a multi-dimensional set of potential influencing factors, including economic, demographic, cultural, and transportation-related variables, and conduct corresponding quantitative analysis. This work complements conventional quality assessment by exploring the underlying mechanisms of quality variations, forming a coherent integrated analytical workflow of multi-dimensional quality evaluation and driving factor attribution. The findings of this study offer empirical insights into the strengths and limitations of crowdsourced highway network data, provide a preliminary empirical reference for its standardized application in transportation research and planning, and may offer supporting references for the development of targeted data quality improvement strategies in open crowdsourced mapping platforms.
During disasters, social media platforms serve as valuable sources of real-time spatial information, capturing the evolving needs of affected individuals and the locations of emergent incidents. However, the unstructured nature of such data, pervasive noise, and the absence of a computational mechanism for translating textual semantics into spatial weights and road network traversal costs leave social media data effectively "visible yet unusable" in disaster analysis and emergency route planning. To address these challenges, this study proposes a social media-driven framework for emergency mapping and risk-aware route planning, using urban flooding as a case study. In the information processing stage, a BERT-TextCNN model is employed for topic classification, filtering approximately 69% of irrelevant content; a BERT-BiLSTM-CRF model is subsequently applied to extract disaster and rescue locations, achieving an F1-score of 95.8% and outperforming the baseline by approximately 4%. In the route planning stage, a multiplicative smooth-decay impedance model transforms discrete disaster events into a continuous spatial risk field, which is integrated with Sort-Tile-Recursive spatial indexing and the Single-Source Shortest Path algorithm to enable real-time, adaptive, hazard-aware navigation across large-scale road networks. Experimental results demonstrate that the proposed framework effectively balances route safety and travel efficiency, offering robust decision support for emergency rescue operations in complex urban environments.
Interferometric Synthetic Aperture Radar (InSAR) serves as a fundamental tool for deformation analysis within Geographic Information System (GIS)-based geohazard monitoring. Phase unwrapping (PU) remains a critical stage for retrieving high-precision topographic and displacement data. The accuracy and efficiency of PU directly dictate the reliability of spatial modeling and emergency responses to geological hazards including landslides and land subsidence. Conventional algorithms typically lack robustness under conditions of intense noise, low-coherence, and steep topography, thereby undermining hazard assessments. To enhance geospatial fidelity and topological consistency, this research introduces and evaluates three independent deep learning (DL) architectures: UA, GL, and CUA. These models leverage U-Net's multi-scale spatial fusion, generative adversarial networks for texture synthesis, and channel attention mechanisms for adaptive feature recalibration. Evaluations on a large-scale InSAR benchmark dataset reveal that these architectures outperform traditional methods in structural fidelity across challenging interferometric conditions. Moreover, they achieve temporal consistency comparable to network flow algorithms in active tectonic scenarios while isolating localized phase errors to halt global propagation, consistently yielding a peak signal-to-noise ratio exceeding 28 dB and a structural similarity index measure above 0.85. These findings provide systematic guidelines for architecture selection in DL-based PU workflows for geohazard monitoring applications.
Accurate and dynamic Intersection Turning Control (ITC) information is a fundamental component for intelligent traffic management. However, widely used open-source maps, such as OpenStreetMap, often suffer from incomplete attribute data and update lags. While crowdsourced trajectory data offers a promising solution, existing extraction methods typically rely on absolute frequency thresholds, lacking a systematic quantification of uncertainty. To address this, we propose an integrated framework for ITC extraction and confidence evaluation using mobile navigation trajectories. The framework first abstracts intersection geometry and refines trajectory segments to ensure spatial alignment. Building on this, a Balanced Random Forest model is utilized to robustly classify movement modes. Crucially, to quantify the reliability of these extracted rules, we introduce a Bayesian hierarchical confidence evaluation model that infers the posterior probability of movement legitimacy from the observed data. This probabilistic approach explicitly quantifies the reliability of each turning rule, effectively distinguishing between rare legal maneuvers and anomalous violations. Experimental results from a case study in Shanghai demonstrate that the proposed framework achieves an overall extraction accuracy of 93.75%, with an optimal confidence threshold identified between 0.60 and 0.65. The study validates that incorporating confidence evaluation significantly enhances the semantic correctness of map updates, providing a robust solution for maintaining real-time, lane-level urban road networks.
Explainable attraction recommendations play a crucial role in enhancing user decision confidence and advancing smart tourism development. In response to the two major challenges in the existing attraction recommendation methods, namely the difficulty in balancing interpretability and recommendation accuracy as well as the lack of transparency in the weight allocation of multi-source heterogeneous information, a method for explainable attraction recommendation that integrates knowledge graph and collaborative filtering is proposed. First, we designed an attraction recommendation metric by comprehensively considering both fundamental attraction characteristics and collective intelligence. Second, we developed an explainable attraction recommendation algorithm that integrates knowledge graph and collaborative filtering. Finally, using attractions in Henan Province, China, as our research area, we conducted accuracy analysis of the recommendation method and designed case studies based on diverse user needs. Results demonstrate that our method achieves 93.8% accuracy, enabling intelligent and effective attraction recommendations while ensuring the explainability of the results.
