DH 2026

Daejeon, July 27–31

Wed, July 2916:30–18:00S077107
Short Paper

Preserving Ukraine’s Urban Cultural Heritage in Crisis: Drone-Based Photogrammetry and AI-Assisted 3D Reconstruction

Iuliia Iashchenko
La Sapienza University of Rome, Italy · iuliia.iashchenko@uniroma1.it
Anatolii Iashchenko
La Sapienza University of Rome, Italy · anatolii.iashchenko@uniroma1.it
Andrea Carteny
La Sapienza University of Rome, Italy · andrea.carteny@uniroma1.it

Abstract:
The ongoing Russo-Ukrainian war has inflicted severe damage on Ukraine’s urban cultural heritage, with Odessa, a UNESCO World Heritage Site, among the most affected. This project proposes an innovative crisis-focused approach to documenting and preserving Ukraine’s historical urban architecture using drone-based photogrammetry combined with artificial intelligence (AI) for data completion and 3D reconstruction. High-resolution aerial imagery captured via drones provides precise spatial data, while AI algorithms fill gaps in incomplete datasets, reconstructing destroyed or partially damaged structures. These 3D models serve both as digital archives and as practical guides for post-war restoration, ensuring the integrity of architectural heritage and supporting cultural resilience amid conflict. By integrating archival resources, historical documentation, and cutting-edge digital tools, the initiative advances the study and preservation of endangered urban environments under conditions of crisis.The ongoing Russo-Ukrainian war has inflicted severe damage on Ukraine’s urban cultural heritage, with Odessa, a UNESCO World Heritage Site, among the most affected. This project proposes an innovative crisis-focused approach to documenting and preserving Ukraine’s historical urban architecture using drone-based photogrammetry combined with artificial intelligence (AI) for data completion and 3D reconstruction. High-resolution aerial imagery captured via drones provides precise spatial data, while AI algorithms fill gaps in incomplete datasets, reconstructing destroyed or partially damaged structures. These 3D models serve both as digital archives and as practical guides for post-war restoration, ensuring the integrity of architectural heritage and supporting cultural resilience amid conflict. By integrating archival resources, historical documentation, and cutting-edge digital tools, the initiative advances the study and preservation of endangered urban environments under conditions of crisis.

Relevance:
Odessa’s historic center, with its eclectic mix of Baroque, Neoclassical, and Art Nouveau architecture, embodies Ukraine’s rich cultural identity. The ongoing conflict poses an immediate threat to these structures, making the development of crisis-oriented preservation strategies urgent. Traditional on-site data collection has become extremely hazardous due to active hostilities, unexploded ordnance, and structurally compromised buildings, placing research teams at significant risk. In this context, drone-assisted photogrammetry emerges as a safe and efficient tool, enabling rapid, high-precision documentation of endangered areas without exposing researchers to life-threatening conditions.

To complement drone-acquired imagery, artificial intelligence (AI) plays a central role in reconstructing missing or damaged visual information. AI-driven workflows can process heterogeneous datasets, including archival photographs, user-contributed images, and social media content, to generate complete, high-resolution visual records suitable for 3D modeling. This combination of drones and AI not only ensures the accuracy and comprehensiveness of digital reconstructions but also provides a practical framework for safeguarding urban heritage during and after armed conflict. By reducing human exposure while maintaining high-quality documentation, this approach supports both the preservation of cultural memory and the planning of post-war restoration efforts, offering a replicable methodology for heritage protection in other conflict-affected urban areas.

Research Design: Leveraging AI and Drone-Based Photogrammetry for Safe Heritage Documentation in Conflict Zones

Research Question:
How can drones and artificial intelligence (AI) be implemented to reduce the fatal risks of on-site data collection in armed conflict while enabling accurate 3D reconstruction of urban cultural heritage through photogrammetry?How can drones and artificial intelligence (AI) be implemented to reduce the fatal risks of on-site data collection in armed conflict while enabling accurate 3D reconstruction of urban cultural heritage through photogrammetry?

Objectives:

  • Document endangered Ukrainian heritage sites using drone-acquired high-resolution imagery.
  • Apply AI techniques to reconstruct missing or damaged visual data, enabling complete 3D models.
  • Integrate historical archives, photographs, and building plans to ensure model authenticity.
  • Foster interdisciplinary collaboration across history, architecture, and digital humanities to develop sustainable crisis-oriented preservation methodologies.

Methodology:
To address these challenges, the project relies on a combination of remote data collection, community engagement, and AI-assisted reconstruction. Historical archives, including photographs, postcards, and building plans, provide essential foundational information. This archival material is supplemented through collaboration with local residents, volunteers, and tourists, who contribute photographs and video footage of Odessa’s architectural heritage. Social media platforms and other online repositories are systematically analyzed to identify additional visual data, providing multiple perspectives and temporal coverage of endangered structures.

Drone-based photogrammetry plays a central role in this methodology, allowing high-resolution aerial imagery to be collected safely from a distance. Fiber drones equipped with stabilized cameras and GPS navigation capture comprehensive views of sites that are either partially accessible or too dangerous for on-site research teams. These drones can maneuver at different altitudes and angles, providing precise spatial data for complex urban environments without exposing researchers to conflict-related hazards.

Once collected, the imagery is processed through AI-assisted workflows to generate accurate and high-resolution data suitable for 3D reconstruction. Photographs sourced from archives, social media, and drones often vary in quality, resolution, and perspective. AI algorithms, trained on architectural patterns and existing photogrammetric datasets, standardize and enhance these heterogeneous images. Missing or damaged sections of buildings can be reconstructed using AI-driven contextual inference, producing complete visual representations even when direct photographs are unavailable. The resulting datasets are then integrated into a comprehensive photogrammetry pipeline, producing detailed 3D models that serve both as digital archives and as actionable references for post-war restoration.

Significance:
This project represents a critical advancement in the preservation of cultural heritage under conditions of armed conflict. By integrating drone-based photogrammetry, AI-assisted image reconstruction, and community-sourced visual data, it enables the systematic, safe, and high-fidelity documentation of Odessa’s historic architecture while minimizing the fatal risks faced by research teams on the ground. Beyond ensuring the accuracy of 3D reconstructions, this approach addresses the urgent need to prevent irreversible loss, creating comprehensive digital records that safeguard cultural memory for future generations.

Moreover, the methodology provides actionable support for post-war restoration, offering detailed 3D models that serve as practical templates for physical rebuilding. At the same time, it fosters global awareness and collaboration, highlighting the international significance of Ukraine’s heritage and promoting shared responsibility for protecting endangered urban sites in conflict zones worldwide. By demonstrating the potential of emerging digital technologies to preserve, reconstruct, and disseminate cultural knowledge even amid crisis, this project establishes a replicable framework for heritage preservation in other war-affected cities and regions, reinforcing both cultural resilience and historical continuity.

Conclusions

By integrating drones, photogrammetry, and AI-driven reconstruction, this project establishes a resilient digital infrastructure for the documentation and preservation of Ukraine’s architectural heritage. It ensures the survival of historical memory and the protection of urban identity amid armed conflict, while providing accurate, high-resolution 3D models to guide post-war restoration. Moving forward, the project plans to expand this methodology to additional endangered sites across Ukraine, incorporate more community-sourced and archival data, and refine AI reconstruction techniques to further enhance model fidelity. In doing so, it aims to create a scalable, replicable framework for safeguarding cultural heritage in conflict zones worldwide, combining technological innovation with crisis-responsive preservation strategies.