EO-VGGT adapts a frozen perspective-oriented 3D foundation model to multi-view optical satellite imagery. Its Geometry-Correlation Constrained Selection (GCCS) selects observations by balancing geometric diversity and radiometric consistency. A Sensor-Ray Encoder (SRE) converts pixel-level pushbroom lines of sight derived from the Rational Function Model into geometry-aware tokens, while a lightweight Ray-Pointing-Aware Adapter (RPAA) injects them into the frozen transformer. The supplied abstract reports no numerical results, datasets, baselines, or ablation measurements.
No heat snapshots are available in the last 24 hours.