- The paper introduces a novel framework that integrates photogrammetry with digital twin technology to monitor the deterioration of historic structures over time.
- Using UAV data and SfM-MVS, the approach generated a dense 3D model with approximately 2.4 million points and validated structural changes via cloud-to-cloud distance analysis.
- The framework enables real-time, non-invasive monitoring that informs timely maintenance decisions and extends applications to broader civil infrastructure.
Monitoring Time-Varying Changes of Historic Structures Through Photogrammetry-Driven Digital Twinning
The research titled "Monitoring Time-Varying Changes of Historic Structures Through Photogrammetry-Driven Digital Twinning" by Xiangxiong Kong presents a framework addressing an under-explored aspect of heritage preservation. The paper proposes a novel integration of photogrammetry techniques and digital twin frameworks to monitor the deterioration of historic structures over time, emphasizing the need for an effective mechanism to track changes rather than static condition assessments.
Overview of the Digital Twin Framework
The paper introduces a digital twin framework composed of five integral components: the real-world physical entity, the virtual entity (a 3D model), the connections for data transfer, the data repository, and a service component that facilitates specific objectives, such as deterioration monitoring. This framework deviates from traditional approaches by removing the data component between the service and physical entity and by utilizing a unidirectional loop for all connections.
The foundation of the framework relies on photogrammetry paired with UAV technologies, utilizing Structure-from-Motion Multi-View-Stereo (SfM-MVS) for reconstructing high-fidelity 3D models of historic structures. By leveraging data from UAVs, high-resolution point clouds are generated to provide precise geometric models of the structures.
Validation and Implementation
Kong's framework was validated using a historic casemate at Fort Soledad in Guam as a testbed. The validation involved several phased processes, including UAV-based image collection, the construction of a virtual dense point cloud entity, monitoring deterioration via cloud-to-cloud (C2C) distance computations, and simulated evaluations of various structural features such as crack-like edges and true cracks.
The UAV operations achieved comprehensive visual captures of the casemate with meticulously planned flight paths, emphasizing points of overlap (70%-80%) for robust data collection. For virtual reconstruction, Agisoft Metashape was employed to process the image data into dense point clouds, with successful segmentation via CloudCompare indicating an impressive precision in the model with approximately 2.4 million 3D points.
In assessing deterioration, the framework calculated C2C distances through point cloud registration and utilized simulation techniques to assess its ability to discern between superficial and structurally significant damage. The result was a method capable of effectively monitoring real-world changes over time.
Practical and Theoretical Implications
Kong's framework exhibits significant potential for diverse applications in cultural heritage preservation and beyond. By facilitating the continuous long-term monitoring of structural conditions, stakeholders can make informed decisions on necessary interventions, optimizing resource allocation and maintenance scheduling. This real-time response capability, although challenging within some heritage contexts, underscores the framework's adaptability and potential for incorporation in other civil infrastructure contexts.
From a theoretical perspective, the integration of photogrammetry with digital twin approaches enriches the existing body of knowledge in both fields. It sets the stage for further developments targeting specific engineering problems such as the non-invasive monitoring of infrastructure elements, which may include detecting microstructural flaws.
Conclusion and Future Directions
The paper concludes with the affirmation that the proposed framework can conduct comprehensive, effective monitoring necessary for the preservation of historic structures. The implications extend into practical scenarios, offering valuable insights for heritage preservation stakeholders. Looking forward, future research directions include the expansion of this framework to broader engineering domains, integrating computer vision techniques for heightened structural health monitoring capabilities.
In summary, the paper offers a robust solution to a critical research gap in the monitoring of structural deteriorations over time, presenting implications for the heritage preservation community and future engineering applications. This work contributes significantly to advancing both photogrammetry-driven methods and digital twin technologies within the domain of cultural heritage.
