Korea Research Institute for Human Settlements: “Discarded Drone Data Should Be Used to Update National Geospatial Information”

[Financial News] High-quality spatial data, frequently generated at construction sites using drones and mobile mapping systems (MMS), are disappearing without being incorporated into national maps. As map currency is directly linked to safety in the era of physical AI—which, like a robot, recognizes real-world spaces and moves on its own—researchers have recommended verifying this data and putting it to use in updating national geospatial information.
According to the Korea Research Institute for Human Settlements on the 27th, Choi Gyeong-a, an associate research fellow at the Center for Spatial Information Policy Research, proposed these policy measures in a Land Policy Brief. The research team examined how spatial data from buildings and road construction sites are produced and managed. It compared these practices with the notification system for precision road maps to identify gaps in the institutional framework.
The research team identified the current update system as a limitation. Because the government locates areas that have changed and conducts new surveys, costs rise as the area requiring updates expands. This approach makes it difficult to keep pace with changes across the country. If maps do not match real-world spaces, AI systems that rely on them to make decisions may malfunction, potentially leading to safety accidents. Meanwhile, precise data are steadily accumulating at construction sites through drones and MMS, but there is no channel for this data to flow into national geospatial information. The team said this has created a “dual structure” in which the government and construction sites operate separately.
For buildings, data captured by drones during construction remain solely in the files of the commissioning agency. Drawings and surveying results submitted for use approval, as well as construction records, are used only for building administration. This means that while information about a building’s attributes is updated, the data are not used to update its geometry to reflect its actual shape. For roads, final public surveying results are submitted to the National Geographic Information Institute (NGII). A system is in place for road management authorities to submit standardized MMS data for precision road maps, but it applies only to sections that have already been mapped.
The report proposed four broad solutions. First, it recommended making source data collected at the time of project completion the top priority for collection. It proposed revising the “Guidelines for Preparing Construction Completion Drawings for Revising Digital Maps” to add source data and a minimum set of metadata to the list of required submissions. It also recommended using existing submission channels rather than creating new ones. The proposal is to incorporate the requirement into established procedures, such as completion notifications, handover of public surveying results and use approval, and to apply it in stages, starting with public buildings. Before accepting the data, the report proposed an initial verification step that would classify it by quality level as “incorporate immediately,” “use after quality improvements,” or “use as a clue for detecting changes.” It also called for quality verification guidelines to be formally established. The report set out a division of responsibilities: the NGII would receive and incorporate the data; the Spatial Information Quality Management Institute would assess its quality; and permitting authorities and local building safety centers would verify building information. It also recommended offering incentives to data providers, such as enabling them to reuse the updated data.
The research team considered this a good time for a transition, as government plans already point in the same direction. The Seventh Basic Plan for National Spatial Information Policy (2023–2027) set out plans to shorten the update cycle, currently 0.5 to 2 years, to near real time and establish an automated update system based on satellite and drone data. The Fourth Basic Plan for the Promotion of the Spatial Information Industry (2026–2030) designated spatial information as core data infrastructure for emerging industries such as digital twins, autonomous driving, robot delivery and AI cities. It also included a basis for distributing privately produced spatial information and automating quality management. Technical obstacles such as variations in data quality have also become manageable through the development of technologies to verify and improve the quality of crowdsourced data. The research team emphasized that the remaining task is to establish institutional channels.
Associate Research Fellow Choi emphasized, “In the age of AI, the competitiveness of national geospatial information depends not on how much new surveying is carried out, but on how well the data already being produced are collected, verified and used. Urgent institutional reforms are needed to turn spatial data from construction sites that would otherwise be discarded into a resource for updating national geospatial information.”

[email protected] Jeong Gyeong-su Reporter