Public defence, Geoinformatics, MSc (Tech) Heikki Kauhanen

Title of the thesis: From photogrammetric system development and data analysis to game engine visualization

Thesis defender: Heikki Kauhanen
Opponent: Professor Arzu Çöltekin, FHNW University of Applied Sciences and Arts Northwestern, Switzerland
Custos: Matti Vaaja, Aalto University School of Engineering, Department of Built Environment

The dissertation illustrates an end-to-end workflow that links photogrammetric system development and quality control with geospatial analysis and real-time, geographically accurate visualization in a modern game engine.

Purpose of the study is to increase the operational value of image-based 3D reconstructions by making quality controlled datasets portable across analytical tools and real-time rendering platforms, improving analysis, communication, and reuse of geospatial products.

The work closes a gap between rigorous geospatial analysis and accessible, high-fidelity visualization. It contributes reproducible, scriptable quality assurance for photogrammetry and demonstrates how game engines can host geospatially faithful “digital twins,” informing research and practice in remote sensing, urban environmental monitoring, and digital twin development.

Main result:

• A practical pipeline that moves geospatial analyses from desktop tools (e.g., QGIS, CloudCompare) into Unreal Engine for real-time, geographically anchored exploration without sacrificing spatial accuracy.

New information:

• A scriptable, iterative quality control method that prunes high-reprojection-error points (~1% per iteration) until residuals fall below half a pixel, enabling repeatable, production-grade reconstructions with traceable quality.
• Integrated visualization of watershed analysis, change detection, and ray traced lighting studies as an interactive digital twin where the need for regular updates is identified for operational maintenance.
• Evidence on modality trade-offs in urban green infrastructure: UAV photogrammetry contributes to above-ground biomass estimation but differs systematically from field data and ALS. Thus, hybrid use is recommended. Acquisition conditions (e.g., leaf-off, wind) affect modalities and tasks such as individual tree detection in different ways.
• Guidance on choosing point clouds versus meshes for visualization and communication.
• Modern game engines enable real time visualization of even relatively heavy geospatial models (tens of millions of polygons).

Applications:

• Urban environmental management and planning: interactive, spatially faithful digital twins for green infrastructure assessment and stakeholder engagement.
• Research and production: portable, quality-controlled photogrammetry pipelines that support reproducibility and collaboration across software ecosystems.
• Education and communication: real-time, geospatial visualization to make complex analyses interpretable to non-specialists.
• Digital twin operations: a pathway to maintain living geospatial models through update mechanisms (conceptually scoped).

Conclusions:

• Scriptable photogrammetric quality control enables repeatable, traceable, production-grade pipelines from acquisition to visualization.
• No single sensing modality is universally superior. UAV photogrammetry and ALS are complementary, and hybrid approaches are advised depending on acquisition conditions.
• Bridging analysis and communication in a game engine expands the utility and audience of geospatial products, supports circular data practices, and provides a route toward operational digital twins for urban environmental applications.

Key words: Photogrammetry, data analysis, game engine, circular economy, 3D reconstruction, built environment, point cloud, mesh, drone

Thesis available for public display 7 days prior to the defence at Aalto University's public display page. 

Contact information: heikki.kauhanen@aalto.fi, +358504701577