Although interest in indoor space modeling is increasing, the quantity of indoor spatial data available is currently very scarce compared to its demand. Many studies have been carried out to acquire indoor spatial information from floorplan images because they are relatively cheap and easy to access. However, existing studies do not take international standards and usability into consideration, they consider only 2D geometry. This study aims to generate basic data that can be converted to indoor spatial information using IndoorGML (Indoor Geography Markup Language) thick wall model or the CityGML (City Geography Markup Language) level of detail 2 by creating vector-formed data while preserving wall thickness. To achieve this, recent Convolutional Neural Networks are used on floorplan images to detect wall and door pixels. Additionally, centerline and corner detection algorithms were applied to convert wall and door images into vector data. In this manner, we obtained high-quality raster segmentation results and reliable vector data with node-edge structure and thickness attributes that enabled the structures of vertical and horizontal wall segments and diagonal walls to be determined with precision. Some of the vector results were converted into CityGML and IndoorGML form and visualized, demonstrating the validity of our work.

A critical problem in mapping data is the frequent updating of large data sets. To solve this problem, the updating of small-scale data based on large-scale data is very effective. Various map generalization techniques, such as simplification, displacement, typification, elimination, and aggregation, must therefore be applied. In this study, we focused on the elimination and aggregation of the building layer, for which each building in a large scale was classified as “0-eliminated,” “1-retained,” or “2-aggregated.” Machine-learning classification algorithms were then used for classifying the buildings. The data of 1:1000 scale and 1:25,000 scale digital maps obtained from the National Geographic Information Institute were used. We applied to these data various machine-learning classification algorithms, including naive Bayes (NB), decision tree (DT), k-nearest neighbor (k-NN), and support vector machine (SVM). The overall accuracies of each algorithm were satisfactory: DT, 88.96%; k-NN, 88.27%; SVM, 87.57%; and NB, 79.50%. Although elimination is a direct part of the proposed process, generalization operations, such as simplification and aggregation of polygons, must still be performed for buildings classified as retained and aggregated. Thus, these algorithms can be used for building classification and can serve as preparatory steps for building generalization.

In recent years, question answering on knowledge bases (KBQA) has emerged as a promising approach for providing unified, user-friendly access to knowledge bases. Nevertheless, existing KBQA systems struggle to answer spatial-related questions, prompting the introduction of geographic knowledge ba se question answering (GeoKBQA) to address such challenges. Current GeoKBQA systems face three primary issues: (1) the limited scale of questions, restricting the effective application of neural networks; (2) reliance on rule-based approaches dependent on predefined templates, resulting in coverage and scalability challenges; and (3) the assumption of the availability of a golden entity, limiting the practicality of GeoKBQA systems. In this work, we aim to address these three critical issues to develop a practical GeoKBQA system. We construct a large-scale, high-quality GeoKBQA dataset and link mentions in the questions to entities in OpenStreetMap using an end-to-end entity-linking method. Additionally, we develop a query generator that translates natural language questions, along with the entities predicted by entity linking into corresponding GeoSPARQL queries. To the best of our knowledge, this work presents the first purely neural-based GeoKBQA system with potential for real-world application.

Large language models (LLMs) have advanced geospatial artificial intelligence; however, geospatial knowledge-base question answering (GeoKBQA) remains underdeveloped. Prior systems have relied on handcrafted rules and have omitted the splitting of datasets into training, validation, and test sets, thereby hindering fair evaluation. To address these gaps, we propose a prompt-based multi-agent LLM framework (based on GPT-4o) that translates natural-language questions into executable GeoSPARQL. The architecture comprises an intent analyzer, multi-grained retrievers that ground concepts and properties in the OSM tagging schema and map geospatial relations to the GeoSPARQL/OGC operator inventory, an operator-aware intermediate representation aligned with SPARQL/GeoSPARQL 1.1, and a query generator. Our approach was evaluated on the GeoKBQA test set using 20 few-shot exemplars per agent. It achieved 85.49 EM (GPT-4o) with less supervision than fine-tuned baselines trained on 3574 instances and substantially outperformed a single-agent GPT-4o prompt. Additionally, we evaluated GPT-4o-mini, which achieved 66.74 EM in a multi-agent configuration versus 47.10 EM with a single agent. The observations showed that the multi-agent gain was higher for the larger model. Our results indicate that, beyond scale, the framework’s structure is important; thus, principled agentic decomposition yields a sample-efficient, execution-faithful path beyond template-centric GeoKBQA under a fair, hold-out evaluation protocol.

Recently, open-domain question-answering systems have achieved tremendous progress because of developments in large language models (LLMs), and have successfully been applied to question-answering (QA) systems, or Chatbots. However, there has been little progress in open-domain question answering in the geographic domain. Existing open-domain question-answering research in the geographic domain relies heavily on rule-based semantic parsing approaches using few data. To develop intelligent GeoQA agents, it is crucial to build QA systems upon datasets that reflect the real users’ needs regarding the geographic domain. Existing studies have analyzed geographic questions using the geographic question corpora Microsoft MAchine Reading Comprehension (MS MARCO), comprising real-world user queries from Bing in terms of structural similarity, which does not discover the users’ interests. Therefore, we aimed to analyze location-related questions in MS MARCO based on semantic similarity, group similar questions into a cluster, and utilize the results to discover the users’ interests in the geographic domain. Using a sentence-embedding-based topic modeling approach to cluster semantically similar questions, we successfully obtained topic models that could gather semantically similar documents into a single cluster. Furthermore, we successfully discovered latent topics within a large collection of questions to guide practical GeoQA systems on relevant questions.

In a situation where cities are increasingly being developed vertically and complexly, a novel approach for analyzing vertical urban functional patterns is proposed. For this purpose, a multi-dimensional GDC (Geo Data Cube) consisting of spatial and temporal data x, y, z, t, and f dimensions containing layer information was created. At this time, the size of the GDC cell (interval in x, y, z dimensions) is calculated by cell point data using the three-dimensional (3D) Moran’s I index value calculated with the 3D Diversity Factor (DF) based on information entropy proposed to reduce the uncertainty of information for each cell. In other words, the cell with the smallest index value was chosen to minimize the influence of Modifiable Areal Unit Problem (MAUP) that occurs when mapping. The 3D land use index (3D LUI) is calculated as a linearly weighted sum of the spatial accessibility of uses between cells (3D KDF) and the enrichment of uses (3D EF), taking into account the first law of geography. Finally, the 3D LUI value for each use was calculated for each cell of the GDC, and the use with the highest value was determined as the urban function of the cell. As a result of applying this to Seocho-gu, Seoul, Republic of Korea (ROK) in June 2024 and visually evaluating it using the street view provided by Kakao Map, it was confirmed that commercial and residential functions were vertically separated in buildings with residential–commercial complexes or shops on the ground floor. It was also confirmed that such characteristics did not appear in the two-dimensional (2D) urban functional patter analysis.