Fusion of BIM and SAR for Innovative Monitoring of Urban Movement – Towards 4D Digital Twin

Abstract. Digital Twin has gained significant attention for future urban management. Since 2020, Building Information Modelling (BIM) has become the standard for high-rise buildings and infrastructures in Germany. Our project BIMSAR is to combine BIM data and SAR-derived movement for urban movement monitoring. The so-called persistent scatterer (PS) points are extracted from a SAR image stack based on PSI. These PS-points, considered pseudo-substructures of buildings, contain geoinformation such as geographic position and movement estimates. They are integrated into the BIM models for various self-contained units like façades and roofs. This fusion step is implemented by machine learning, employing a novel distance metric adapted through dimensionality reduction. To provide a user-friendly analysis tool, the fused data are displayed in 4D (position + movement) on a GIS-based platform, along with other geodata, e.g., geological maps, hydrological maps, underground mining structures. This paper shows the current results around Ahlen in Ruhr region, where many damages to buildings, bridges, and highways are associated with the abnormal deformation. We created the BIM models ourselves by converting open-source 3D models, adhering to Industry Foundation Classes (IFC) format. The SAR data contain Sentinel-1, TerraSAR-X, and PALSAR-2 for a cross-analysis. The primary application is to assist stakeholders with an early warning system to evaluate whether unusual or long-term deformations cause structural damages and thereby threaten the safety of life and property. We demonstrate some examples towards this application.