International Conference on Geoinformatics : [proceedings]

Exploring the public's cognitive status, participation level, and emotional tendency towards intangible cultural heritage (ICH) is conducive to gaining insights into the public's attitudes and needs towards traditional culture, and provides a key basis for accurately formulating intangible cultural heritage protection and inheritance strategies. This paper crawls ICH text data from microblogs, adopts word frequency analysis, theme mining, emotion classification and other methods to excavate the deeper meaning of the text, and explores the public's real concerns, emotional attitudes and their formation mechanisms, and the results show that: 1) the public's perception of the ICH presents a multi-dimensional feature, and through theme analysis, it is found that the public discussion mainly focuses on three dimensions: cultural inheritance and development (52%), implementation and experience of the activities (32%), and introduction and promotion of ICH (16%). Among them, the proportion of topics related to the introduction and promotion of ICH is relatively low in microblogging, and it is suggested that the relevant departments increase the publicity and promotion of specific ICH. 2) The results of the sentiment analysis show that positive sentiments dominate the public's sentiment towards ICH. The average value of sentiment in each region exceeds 0.67 (out of 1), with Jiujiang City (0.8) and Ganzhou City (0.79) significantly leading the way. 3) Spatial econometric analysis reveals significant regional differences in the distribution of ICH sentiment. This difference was jointly influenced by multiple factors, of which the number of ICHs, GDP, and education expenditure level were the three most explanatory influences, with regression coefficients of 0.911,0.906, and 0.897 (p<0.01), respectively. This finding provides an important reference for the development of differentiated ICH protection policies.

Since 2021, the “low-altitude economy” has been successively incorporated into national plans and government work reports, emerging as a strategic emerging industry. As an economic form driven by low-altitude airspace, its development core lies in transportation, and the key to breakthroughs is traffic management. This paper sorts out the connotations of low-altitude economy, transportation, and traffic management, analyzes the elements of low-altitude economic development and traffic management, and explores the management forms of low-altitude manned aircraft, unmanned aerial vehicles (UAVs), and urban air mobility. The study finds that it is necessary to break through traditional management bottlenecks with a “low-altitude mindset”, and build a traffic management system adapted to the development of the low-altitude economy through mechanism innovation, regulation improvement, and talent cultivation, so as to promote the low-altitude economy to become a new growth pole of the national economy.

As the core growth pole of the Central Plains Urban Agglomeration, Zhengzhou's rapid urbanization has significantly disrupted regional ecological security patterns. This study integrates multi-temporal Landsat remote sensing data (2000-2020) to develop an Improved Remote Sensing Ecological Index (IRSEI) that synthesizes fractional vegetation cover (FVC), normalized difference built-up index (NDBI), and land surface temperature (LST). Combining a Geographically and Temporally Weighted Regression (GTWR) model, we systematically quantify the ecological response mechanisms to urban expansion across 12 counties/districts in Zhengzhou. Results reveal three key insights: First, spatial patterns exhibit “monocentric dominance-gradient differentiation,” with the southeastern plains (Zhongmu County, Xinzheng City) experiencing an annual expansion intensity of $\text{8. 2 1 \%}$, driving increases of 35.6% in patch density and 59.4% in landscape shape index, while the northwestern ecological barrier zone (Dengfeng City, Xingyang City) maintained a maximum patch index above 0.62 under ecological conservation redline constraints. Second, IRSEI evolution displays a “rise-then-decline” trend, with its mean value fluctuating from 0.469 (2000) to 0.487 (2020). Ecological quality improved by 12.6% in northwestern mountainous areas, whereas the southeastern core zone (Jinshui District, Zhongyuan District) suffered a 7.3% IRSEI decline due to synergistic heat island effects and construction land sprawl. Third, GTWR modeling identifies expansion intensity and landscape morphology as dominant drivers, where a 1-unit increase in southeastern urban expansion intensity reduces IRSEI by 0.38, with landscape fragmentation explaining 68% of ecological degradation. This research provides a quantitative basis for balancing territorial spatial development and ecological conservation in the Zhengzhou Metropolitan Area, and the IRSEI-GTWR framework offers a novel methodology for sustainable urban agglomeration assessments.

With the continuous expansion of highway networks in recent years, the monitoring and repair of road diseases have become one of the important tasks in traffic management. Traditional manual inspection methods are inefficient and prone to missing disease areas. Therefore, a pavement disease detection system based on UAV vision and deep learning algorithms has emerged. The formation mechanism of pavement diseases was first analyzed, and the UAV pavement image acquisition equipment was introduced. Then, a pavement disease identification algorithm using YOLOv8 was constructed. Finally, Python programming was employed to complete the construction of a regional identification model for highway pavement diseases based on CNN, data loading, and model training and testing. The experimental results show that the system exhibits high accuracy and real-time performance in image, video, and real-time camera identification, effectively assisting in the monitoring of highway diseases. This system can significantly improve the automation level of pavement disease detection, reduce manual workload, enhance road management efficiency, and provide strong support for the safety and maintenance of highways.

Coastal cities are among the fastest urbanizing and industrializing regions, but they also face critical issues, including environmental degradation, uncontrolled urban expansion, and an intense increase in carbon discharges. Hence, it was necessary to conduct a rigorous assessment of changes in landscape coverage and carbon storage. This research focused on Lianyungang and proposed an integrated method, CMCPI, which coupled multi-feature for land classification, PLUS, and InVEST. The purpose was to detect alterations in land utilization patterns and evaluate carbon discharges. Utilizing remote sensing data from 2010 to 2020, a detailed land use classification was performed for the research area, and its spatiotemporal change characteristics were analyzed. Subsequently, simulations predicted land use patterns for 2030 across four distinct scenarios: Nature-Focused Development (ND), Urban Rapid Development (UD), Cropland Protection (CP), and Ecological Sustainable Protection (EP). Additionally, the variations in regional carbon storage across different scenarios were assessed. The consequences indicated that the CMCPI achieved high accuracy and reliability in land classification. From 2010 to 2020, impervious surface areas in Lianyungang significantly increased to 1276.62 km2, while cropland, forestland, and grassland areas decreased. Carbon storage initially declined and then increased. Predictions for future scenarios indicated that only the EP scenario would promote an increase in carbon storage, reaching 9200.25 mmt, whereas the other three scenarios would result in varying degrees of carbon storage reduction. This research not only provides new insights into understanding the ecosystem benefits service values and carbon storage capacities of coastal areas but also offers scientific support for policy-making towards achieving “carbon neutrality” and provides references for sustainable development planning for Lianyungang.

Photovoltaic (PV) power generation is becoming increasingly important as a clean, renewable form of energy, and accurately identifying and calculating the area of PV panels is critical to the planning, monitoring, and optimization of PV systems. Remote sensing technology, due to its large-scale, cyclical, and non-contact data acquisition capabilities, can efficiently obtain imagery of PV power plants across extensive areas, thereby establishing a comprehensive database for PV panel identification. However, in practical applications, intelligent identification of PV power plants still faces several challenges. Individual elongated PV panels with low spatial distinguishability are difficult to recognize, while planned undeveloped areas in planning stages exhibit spatial signatures closely resembling built-up regions, posing challenges for effective discrimination based solely on spatial attributes. Therefore, to address the above challenges, this paper proposes the Spectral-High Resolution Network (S-HRNet), based on HRNet and incorporating a spectral feature extraction module. The model is validated using the Taratan PV power plant case study. The model extracts spectral information from images to enhance existing feature representation while optimizing the feature fusion mechanism, enabling accurate PV panel recognition through the synergy of spatial-spectral features. The model demonstrated effectiveness in the Taratan PV plant dataset, outperforming HRNet by achieving improvements of 1.47 % and 0.79 % in IoU and F1, respectively, and notably reaching 95.28 % IoU and $97.34 \% ~\mathrm{F} 1$. When applied to the Taratan PV plant, the model identified a photovoltaic panel area of $251.42 ~\text{km}^{2}$, outperforming HRNet's $252.52 ~\text{km}^{2}$ and closely matching the visually interpreted $251.77 ~\text{km}^{2}$. This result verifies the model's practical value in large-scale PV plant monitoring and provides reliable technical support for applications such as dynamic PV resource assessment and power generation efficiency prediction.

The explosive technological iterations in the current field of artificial intelligence have provided new ways of thinking and technical means for the analysis of spatial big data. Geographic Information Artificial Intelligence (GeoAI) is developing rapidly and has become a hot field of current research. This paper provides a detailed review of the cutting-edge research in GeoAI. Firstly, it clearly defines the scope of GeoAI research in the fields of geographic science and urban spatial science, and categorizes it into three types: symbolic GeoAI, connectionist GeoAI, and behaviorist GeoAI. Then, by introducing the concept of “semantics” and based on the semantic types extracted by models, it further classifies these three logical approaches of GeoAI research, aiming to deeply explore the research questions, content, and technical methods involved in each classification. On this basis, the article summarizes current research focuses and proposes potentially promising future research directions in the era of large models. Through the study, it is found that symbolic GeoAI can effectively organize static geographic knowledge, dynamic geographic processes, and spatio-temporal urban knowledge; connectionist GeoAI focuses on spatial embedding-based semantic representation and prediction; while behaviorist GeoAI concentrates on assisting spatial plan generation. In practical applications, there are the largest number of connectionist GeoAI studies, which are often used in combination with the other two types of GeoAI in upstream and downstream processes. Overall, GeoAI is increasingly pursuing the integration of multimodal data sources; different types of GeoAI research may share common goals, increasingly requiring the combination of multiple modeling logics to solve problems; and the flexible use of large models as intelligent agents to help solve geographic problems has become a new feasible path.

Regional population mobility exhibits dynamic adjustments in response to external environmental shifts. This dynamic process reflects the interaction of demographic flows with economic resilience. By using Baidu Migration data and statistical yearbook records, this study investigates the spatiotemporal patterns during China's Spring Festival population movement before and after the COVID-19 outbreak (2019-2020) through spatial autocorrelation analysis and geographical detector methods. The results show three key findings. (1) The scale of population movement fluctuates significantly around the Spring Festival period, with migration volumes gradually recovering and stabilizing during the post-outbreak phase, highlighting mobility's resilience to shocks. (2) Distinct regional patterns emerge in population movement. Intra-provincial and inter-provincial migration exhibits clear east-west and north-south divergences across prefecture-level cities. Four major urban agglomerations serve as primary population concentration centers. These patterns illustrate the dynamic interplay between population distribution and economic development, as well as their resilience to external changes. (3) Economic factors significantly influence migration forms and regional disparities. Social consumption is the key driver of differences in intra-provincial, inter-provincial, and regional (east-central-west) migration. The interaction among factors mainly shows a two-factor enhancement effect. These findings reveal the economic system's underlying resilience mechanisms in supporting mobility changes.

The sudden occurrence of flood-related natural disasters poses a severe threat to human life and property safety. Rapid and accurate extraction of building damage information is crucial for humanitarian aid and emergency response. Flooddamaged building information extraction relies on background information; however, existing algorithms for extracting flooddamaged building information predominantly focus on foreground features, resulting in incomplete damage category recognition and a high rate of false alarms. To address this issue, this paper proposes a Flood-damaged Building-Net (FDB-Net) algorithm based on Local-Global Collaborative Attention. The algorithm extracts features through foreground and background attention networks, and subsequently captures local and global features through parallel convolution and Transformer networks, thereby enhancing the network's ability to perceive foregroundbackground relationships and model long-distance dependencies. In the decoder, by modeling the relationship between dualtemporal features, the algorithm effectively integrates semantic and spatial information from multi-temporal images, combining spatial features with temporal change characteristics to alleviate false detection and missed detection issues. Through feature interaction and collaborative attention, the model increases sensitivity to flood-damaged building areas while suppressing irrelevant or redundant information. Experimental validation on the xBD-Flooded dataset demonstrates the effectiveness and advantages of FDB-Net, showing significant improvements over several existing building damage extraction methods.

Developing factories in rural areas suited to their local conditions is very important for rural revitalization and integrated urban-rural development. Ecological and environmentally friendly efficient land use of rural areas play a key role in a regional sustainable development, which needs the information of land use of rural factories in detail in order to make their land use ecological and efficient. It is difficult to obtain the land use information in detail of rural factories because their distribution is extremely scattered. In order to overcome the difficulty, a new methodology is developed here. Several indicators proposed here which are used to describe the land use of rural factories in detail include building density, plot ratio, vegetation coverage rate, cement pavement coverage rate and water coverage rate. The methodology includes: (1) obtaining images of a rural factory by using Dajiang unmanned aerial vehicle (Dajiang UAV); (2) processing the images with camera lens model, producing ortho-images and mosaicking images; (3) extracting the area of a rural factory, the number of floors of buildings, vegetation area, cement pavement area and waterbody area in the rural factory from the UAV images, and creating their database; (4) calculating the indicators by using the database. A rural factory in Huanglong village of Chengdu plain, located in Qingliu Town of Xindu District of Chengdu city, is used as an example to explore the methodology here. The research shows that the methodology is of low cost, high efficiency and accuracy, and easy to be grasped, which has a broad application value in quickly and accurately obtaining the land use information in detail of rural factories which are sparsely scattered in vast rural area. The land use information in detail is essential for rural revitalization and construction of beautiful China. The research also discovered that the rural factory in Huanglong village is of 0.378 of building density, 0.435 of plot ratio, 23.69 % of vegetation coverage rate, 28.59 % of cement pavement coverage rate and 2.83 % of waterbody coverage rate.

As cities continue to densify and high-rise developments become more prevalent, residents' views of their surroundings are often obstructed, limiting their visual exposure to natural features such as greenery. Given the increasing amount of time individuals spend indoors, assessing their green views from buildings has become essential, as greenery contributes to livability and well-being. This study develops an observer-centric measure to quantify individual green views from apartment buildings using a spatially explicit framework. The proposed method leverages GIS-based 3D techniques and spatial analysis to capture and quantify green views. Additionally, grey views (built-up environments) and water views (surface water) were quantified to account for the mix of natural and urban features within an observer's views. The framework was tested on sample buildings in Melbourne, Australia, using high-resolution environmental datasets and 3D building models. The results demonstrate significant variations in observers' green views based on building type, floor level, and surrounding urban form. The approach provides valuable insights for urban planners, architects, and policymakers, enabling data-driven strategies to optimize greenery access, enhance urban design, and promote healthier, more sustainable cities. Additionally, it supports health-related policies by facilitating the assessment of green exposure's impact on wellbeing, informing interventions to improve mental and physical health in urban environments.

This research focuses on the precise identification and distribution characteristics of oases in the arid regions of Northwest China, developing a process utilizing ArcGIS spatial analysis and manual refinement. Utilizing data from the Fifth National Desertification Monitoring, as well as geographic national conditions monitoring Land Cover and hydrological data, the study extracts and analyzes oases within a 100kilometer buffer zone around major deserts. The results indicate that oases are primarily distributed in bands or sporadically around deserts, correlating with regional water resource supply, topographical conditions, and human activities. The study also finds that oasis in the Qinghai-Tibet Plateau are predominantly covered by forest and grassland, while deserts and barren lands are concentrated in the Qaidam Basin; the Ningxia and Hetao plains feature a mosaic of cultivated land and residential areas. This research provides a scientific basis for ecological evolution analysis and sustainable development of oases, offering technical support to address climate change and anthropogenic impacts.

Few-shot learning has important applications in remote sensing image classification due to its strong generalizability. The objective of few-shot learning is to leverage prior knowledge to address new tasks with only a small number of fine-tuning samples. Therefore, How to accumulate prior knowledge during training stage is a key problem in solving few-shot classification problem. In addition, most existing methods suffer from the class confusion because of the inter-class similarity and intraclass diversity of remote sensing images. To alleviate these problems, this paper proposes Hybrid Learning with MultiObjective Optimization (HL-MOO) to enhance few-shot remote sensing scene classification performance. Specifically, in training stage, we propose a joint learning strategy that combines two selfsupervised learning tasks and a classification task to capture the intrinsic features of remote sensing images. We also introduce a multi-objective optimization algorithm to dynamically adjust the impact of each task on our model. HL-MOO not only improves the performance on training samples, but also enhance adaptability to new scenes. It is evaluated on NWPU-RESISC45, Aerial Image Dataset (AID), and WHU-RS19 benchmarks, achieving average accuracies of $\mathbf{8 1. 0 9 \%, ~} \mathbf{7 8. 4 0 \%}$, and $\mathbf{9 6. 3 4 \%}$ in the 1-shot setting, and $89.58 \%, 88.12 \%$, and 97.89 % in the 5 -shot setting, respectively. We also test its cross-domain performance on UCM dataset. Results show that our HL-MOO exhibits improvements in few-shot remote sensing scene classification tasks compared to state-of-the-art models.

Since the early 2000s, Ashburn, a satellite city to DC, has urbanized rapidly and has turned into a “Data Center Alley.” Approximately 70 % of internet traffic passes through the region. The data centers have led to an increase in the Urban Heat Island (UHI) effect, which impacts the local environment. This study aims to quantify the extent of the UHI effect and the key causes of the side effects. Satellite data was analyzed through various statistical methods and comparisons and various regions of interest were considered. Additionally, machine learning techniques were applied for downscaling the Land Surface Temperature (LST) data for the purpose of understanding the temperature distribution in a finer detail. The analysis yields that there has been a significant increase in LST, particularly in areas near data centers. This research highlights the urgent need to regulate urbanization and data centers for environmental sustainability. The environmental impact of recent technological innovation and urban development has raised temperatures and destroyed crucial green land that helps combat climate change.

Urban lakes play a crucial role in mitigating the urban heat island (UHI) effect, but the relationship between lake morphology and cooling efficiency remains not fully understood. Addressing this research gap, our study focuses on the lakes in Wuhan, integrating remote sensing technology and data analysis to investigate the influence of lake morphology on cooling capacity. The research emphasizes three main aspects: a) developing a methodological framework to estimate the spatial extent of lake cooling influence with different sizes; b) sing partial correlation analysis to reduce interference from other factors, and analyzing the correlation between lake morphology indices and cooling efficiency, with particular attention to spatial variations between central and non-central urban areas; and c) exploring the mechanisms through which lake morphology influences cooling capacity. The findings reveal that in central urban areas, complex shorelines and higher fractal dimensions significantly improve cooling efficiency and capacity, while in non-central areas, the correlation is weaker. This study not only advances the scientific understanding of the ecological functions of urban lakes but also provides practical insights for urban planning and ecological construction, offering solutions to alleviate urban heat islands and create more livable and sustainable urban environments.

The Qinhe River Basin is a crucial ecological barrier and water conservation area in the Yellow River Basin, playing an indispensable role in supporting national ecological protection strategies and promoting regional high-quality development. An in-depth study of its hydrological processes holds significant scientific importance and can provide key insights for sustainable management. This research investigates the coupled impact mechanisms of climate change and land use/cover change (LUCC) on runoff processes. Using meteorological data from 1979 to 2018, combined with land use data from 1990 to 2020, a multi-scenario simulation method was employed to quantitatively evaluate the contributions of precipitation, temperature changes, and LUCC to runoff. The results indicate that: (1) The SWAT model has high applicability in the Qinhe River Basin, with Nash-Sutcliffe Efficiency (NSE) and $\mathbf{R}^{\mathbf{2}}$ values both greater than $\mathbf{0. 7 5}$ during the calibration and validation periods, and the percentage bias (PBIAS) less than 10%, indicating excellent simulation accuracy; (2) Precipitation is the dominant driver of runoff change, with a 10 % increase or decrease in precipitation leading to a $17.5 \%-25.4 \%$ change in runoff. Based on the baseline scenario under constant precipitation conditions, a 1° C temperature increase reduces runoff by 3 %, while a 1° C temperature decrease increases runoff by 3.4 %; (3) From 1990 to 2020, LUCC in the study area mainly manifested as the conversion of cultivated land to builtup land (a 106.1 % expansion), which increased the impervious surface, resulting in an average monthly runoff increase of $\mathbf{0. 6 6} \mathbf{m}^{3} \boldsymbol{/} \mathbf{s}$ in 2010 compared to 2000, and exacerbated the risk of extreme runoff events. Based on a comprehensive analysis, climate change is the primary factor influencing runoff variation in the Qinhe River Basin, while the impact of Land Use and Cover Change (LUCC) is relatively minor, though it exhibits cumulative effects.

The increasing popularity of geographically weighted (GW) techniques has resulted in the development of several software packages. Their ongoing updates and enhancements always require extraordinary efforts. In this study, we introduced a fundamental C++ library, namely libgwmodel. With its updates and maintenance in the future, all the associated products could be uniformly maintained and upgraded. Moreover, its development will greatly facilitate the further achievement of more GW models or tools, even from different teams, so as to form a comprehensive foundation for further development and extensions of GW models.

Driven by global climate change and the “0Dual Carbon” goals, the precise analysis of spatiotemporal carbon sourcesink dynamics has become a central challenge in environmental governance. Traditional approaches relying on single data sources and static models often fail to adequately capture spatiotemporal heterogeneity. To address this, this study proposes a multi-source sensing collaborative-driven knowledge graph construction method for carbon source-sink dynamics. The framework integrates “space-air-ground-social” multi-source data to build a spatiotemporally continuous ontology model, featuring a four-dimensional representation system (semantic, spatial, temporal, and attribute features). It further incorporates domain knowledge-enhanced large language models to enable efficient extraction of unstructured text and structured data. Experimental results demonstrate that the method successfully achieves intelligent extraction of carbon-related entities and relationships from unstructured text, with the entity recognition accuracy of 92.3 %. A city-scale, dynamically updated carbon source-sink knowledge graph was constructed for Shenzhen, providing granular decision support for regional carbon sequestration potential assessment and emissions trading. This work advances carbon governance from an experience-driven to a data-driven paradigm.

Maize canopy height (CH) monitoring directly reflects maize growth status and potential yield, but is cumbersome through manual in-field measurements. Unmanned Aerial Vehicle (UAV) remote sensing has made this task highly effective. However, the optimal variable to represent maize CH varies in different studies. In this study, we confirmed and compared the feasibility of using UAV-borne LiDAR and RGB to estimate maize CH in multiple treatments and growth stages. The optimal variables for estimating CH were determined for RGB and LiDAR, respectively. The optimal combination (LiDAR, 99th percentile of the digital surface model, minimum digital terrain model) proved robust when applied to the independent data in 2021 (R2: 0.972 RMSE: 0.233 m, rRMSE: 19. 1 %). The proposed non - parametric and straightforward method for maize CH monitoring can be easily transferred to other crops, and has the potential to be applied as a common phenotyping procedure for precision agriculture and smart breeding.

Segmentation of tubular structures in remote sensing imagery represents a domain of significant value for geographic information systems. However, achieving high-quality automated segmentation remains challenging due to morphological diversity and boundary ambiguity between tubular structures and their backgrounds. To address these challenges, we propose RLRAnet, an enhanced Unet-based segmentation architecture that incorporates boundary-aware mechanisms to improve tubular structure extraction. Specifically, we designed a reverse Attention Module that performs inverse calibration on low-level features within multi-scale skip connections, thereby enhancing boundary detail representation. Additionally, we introduced a reinforcement learning-based dynamic loss weight adjustment strategy that leverages a policy network to adaptively balance Soft Dice loss and boundary loss, achieving optimal reconciliation between global segmentation and local detail preservation. Extensive evaluations on two distinct remote sensing tubular structure datasets demonstrate that RLRAnet significantly outperforms multiple advanced segmentation architectures across IoU, Accuracy, Recall, and F1 Score, substantiating its superior performance in tubular structure segmentation tasks.