Volumetrische Videos zur Datenerhebung: Laborbericht einer spielerischen Annäherung im GameLab

The implementation of a ten-camera Microsoft Azure Kinect DK system at the University of Konstanz’s GameLab serves as a foundational exploration into capturing volumetric video for the MARS-D research project. By integrating these spatial interaction data into the Unity game engine, the project establishes a functional workflow for creating immersive, three-dimensional learning environments. The primary objective is to evaluate the practical feasibility of using volumetric capture to simulate complex classroom scenarios, providing a technical framework for future pedagogical research.

The technical architecture relies on Depthkit software, which proves highly effective for processing volumetric data despite its significant licensing costs. The calibration process, facilitated by QR markers, allows for precise spatial alignment within virtual environments, enabling users without advanced programming expertise to develop sophisticated simulations. This workflow successfully bridges the gap between raw sensor data and interactive virtual reality, demonstrating that high-fidelity digital representations of human behavior are increasingly accessible for academic research purposes.

While the current prototype validates the potential of volumetric video in educational settings, several critical challenges remain for future development. Technical limitations persist regarding the capture of small objects and intricate movements, necessitating further refinement in post-production techniques. Furthermore, the project highlights the urgent need for integrated annotation tools within virtual reality and raises essential ethical considerations regarding the creation and storage of digital twins. Future efforts will focus on addressing these technical constraints and establishing robust methodological standards for utilizing volumetric data within broader research contexts.