EO3D

Inspiration

The project will create a collaborative 3D Explosive Ordnance (EO) reference database to improve EO identification. This database will integrate diverse data sources, such as 3D models, technical metadata, and images, using a data fusion approach. Users will be able to interact with comprehensive, multi-dimensional EO profiles. The database will host validated content, including high-quality 3D models and detailed metadata, optimized for searchability, accuracy, and usability. It will also serve as a foundation for future advancements, such as image recognition tools, enhanced training materials, and interoperability with global mine action systems.

Innovation

The project is innovative for several reasons. In particular, it combines:

• AI-enhanced analysis: Automates feature extraction and enriches legacy data.
• Data fusion across modalities: Combines surface scans, volumetric data, and material info into cohesive 3D models.
• Scalable cloud infrastructure: Supports high-performance computing for real-time 3D reconstruction and analysis.

LIST is in charge of creating intuitive, game-style interfaces for EO identification and developing AI tools for object detection and supervised reinforcement learning to integrate human feedback.

Impact

This project will deliver an open-source framework for the 3D digitization and AI-powered identification of explosive ordnance (EO), boosting replicability, enabling global collaboration and embracing the human-centered AI concept. By fusing diverse data types into a unified platform, it empowers faster, safer decision-making in humanitarian demining. Our main partner, the Geneva International Centre for Humanitarian Demining (GICHD), brings deep field expertise and operational insight, ensuring the solution is grounded in real-world needs.