Intelligent 3D monitoring of historical structures with digital twin

Abstract

The preservation of large industrial heritage structures necessitates the implementation of efficient and accessible monitoring techniques to detect structural deterioration and surface changes over time. Conventional monitoring methodologies are predicated on the manual observation of objects, a process that is both highly specialized and time-consuming. The present study proposes a cost-effective and readily transferable framework for multi-temporal 3D change detection using unmanned aerial vehicles (UAVs) and open-source photogrammetry software, with a particular emphasis on machine learning-based difference detection. The framework utilizes structure-from-motion (SfM) and multi-view stereo (MVS) techniques to generate and accurately co-register temporally separated 3D models of technical-historical structures. In contradistinction to conventional approaches, the proposed method is entirely self-referential, relying solely on optical and geometrical features for alignment, thereby eliminating the necessity for specialized surveying equipment. The approach under discussion involves the fusion of both temporal model versions into a composite representation. This model is then used to identify and segment altered regions, thereby distinguishing structural changes such as corrosion.

This approach enables a highly automatable and scalable monitoring pipeline that minimizes the need for manual inspection while increasing the reliability of change detection. By incorporating texture-based comparisons, the framework improves the spatial precision of change localization. Built entirely on open-source tools and lightweight UAV platforms, the proposed method offers a cost-effective and adaptable solution that is especially suited to heritage conservation contexts, where budgetary and operational limitations frequently impede the deployment of conventional monitoring technologies.