Sparse event responses generate uneven keypoint errors, and the impact of these errors on spacecraft pose depends on correspondence geometry and the reliability of successive observations. We introduce a reliability‑aware framework that links polarity‑time localization, geometric recovery, event‑CAD refinement, and trajectory optimization.
The framework uses a three‑branch representation that combines positive events, negative events, and recent‑event timing via cross‑branch attention and gating, followed by spatiotemporal aggregation.
An ordered semantic‑point head estimates relative correspondence precision to enable geometrically constrained pose recovery. Agreement between event edges and CAD projections then guides local pose refinement under keypoint‑geometry safeguards.
At the sequence level, a ten‑dimensional descriptor merges heatmap shape, reprojection consistency, and geometric conditioning to set time‑aligned observation scales.
On the official SPADES challenge test set, the full system achieves a combined pose score of 0.0474, about 5.4% lower than PASAN+IRIS. Validation ablations reduce the score from 0.0772 for base localization with uniform PnP to 0.0514 for the complete cascade; with identical single‑frame observations, quality weighting lowers the score from 0.0549 with constant‑weight optimization to 0.0514.
Review