Reproduce “Automatic Stockpile Volume Monitoring using Multi-view Stereo from SkySat Imagery” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Roger Marí, Carlo de Franchis, Enric Meinhardt-Llopis, Gabriele Facciolo
An automatic system for monitoring industrial stockpile volumes using multi-view stereo from SkySat imagery is presented. The system addresses inconsistencies in camera models for dynamic scenes, achieving accurate 3D reconstructions and demonstrating a strong correlation between satellite-derived volumes and on-site weight measurements.
Key insights
- Relative RPC correction through bundle adjustment significantly improves the consistency of 3D reconstructions from fragmented SkySat imagery, reducing height deviations from meters to tens of centimeters.
- A date-wise processing approach for RPC refinement is critical for accurately monitoring highly dynamic environments where the scene is constantly changing without assuming static terrain.
- Leveraging widely available SmallSat imagery with robust processing techniques provides a scalable and frequent alternative to traditional, often constrained, methods for industrial volume monitoring.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.