robotics
The Localization Matches Became a Depth Sensor
RIDE combined sparse PnP correspondences with a video-depth prior and carried metric scale through short observation gaps.
Summary
RIDE combined sparse PnP correspondences with a video-depth prior and carried metric scale through short observation gaps.
Render–match–PnP relocalization normally uses image-to-map correspondences only to recover camera pose. RIDE reuses the inlier geometry as sparse metric-depth observations, then combines them with a pretrained video-depth prior through global and local correction plus temporal memory. Once a metrically scaled 3D Gaussian Splatting map initializes scale, the system can estimate dense depth from an RGB stream and bridge short periods without new observations. Tests on robot sequences without fine-tuning improved depth accuracy and temporal consistency over scale-only calibration.
Why it matters
RIDE combined sparse PnP correspondences with a video-depth prior and carried metric scale through short observation gaps.
Limits and context
- Render–match–PnP relocalization normally uses image-to-map correspondences only to recover camera pose.
- Tests on robot sequences without fine-tuning improved depth accuracy and temporal consistency over scale-only calibration.
Key claims
RIDE combined sparse PnP correspondences with a video-depth prior and carried metric scale through short observation gaps.
Qualification: Render–match–PnP relocalization normally uses image-to-map correspondences only to recover camera pose.
Evidence: source-2026-09-13-011
Sources
- arXiv preprint 2609.11079arXiv · primary research
Corrections
No corrections have been recorded for this story.