Dabeeo Wins Order from Korea Asset Management Corporation to Detect Changes in State-Owned Property Across South Korea Using AI Technology
Dabeeo, a spatial information AI technology company, announced on the 10th that it has won an order from the Korea Asset Management Corporation (KAMCO) to build an AI-based aerial image change detection system to investigate the status of national property. Dabeeo stated that through this order, it has once again demonstrated its capability to carry out large-scale public projects while validating its spatial intelligence technology.
The project aims to improve the efficiency of nationwide property investigations by reducing the burden of repeated annual field surveys through AI-based aerial image change detection technology. It also seeks to enhance accuracy and efficiency by focusing field surveys only on areas where changes have occurred. In particular, advanced AI spatial analysis is essential, as the system must accurately detect actual changes in land use, overcoming the limitations of seasonality and crop conditions inherent in existing change detection methods. Dabeeo will establish a system that compares multiple image datasets taken from different viewpoints, using 25cm-resolution imagery provided by the National Geographic Information Service. The system is expected to detect various types of changes in state-owned property, such as new construction, unauthorized cultivation, or illegal use. Confirmed changes can then be classified and reflected in property management.
The scope of the project includes training change detection AI models, developing inference systems, building high-performance infrastructure to support nationwide change detection within 90 days, and linking the results with existing systems such as KAMCO’s “National iN-GIS.” Dabeeo plans to improve both precision and versatility by comparing semantic differences between images using a Siamese-like network structure, rather than relying solely on classification-based methods. The goal is to achieve more than 90% accuracy in change detection based on the F1-score standard.
In addition, the project is expected to accelerate the digital transformation of state-owned property management by visualizing detection results through a GIS-based intuitive interface and establishing a feedback structure that enables continuous reinforcement learning. “This project is a leading example of AI-based change detection technology being applied to land management systems, and it will serve as an important milestone in advancing digital infrastructure in the public sector,” said Park Joo-heum, CEO of Dabeeo. “We will continue to contribute to improving the efficiency of public administration through the advancement of spatial information technology.”
댓글
댓글 쓰기