Driving · 2026-11-11
GEMOR-GS: Global Extrinsic-calibration-refined Mapping and Online Relocalization for 3D Gaussian-Splatting SLAM
Yuchen Yang, Ming Ding, Xinyu Wu, Hesheng Wang
AI-assisted summary based on the paper. Consult the linked source for methods and limitations. Benchmark numbers appear after editor verification.
TL;DR
GEMOR-GS is a LiDAR–inertial–visual SLAM system that upgrades a 3D Gaussian splatting (3DGS) map into a geometrically accurate, globally refined, and re-enterable world representation. It estimates LiDAR–camera extrinsics offline via direct-method alignment and reuses them. Offline refinement alternates gated differentiable rendering with LiDAR-projection objectives, followed by global reconstruction with online constraints lifted. For tracking loss, it relocalizes via an XFeat bypass map, RANSAC-PnP, and Gaussian RGB-D warp refinement.
Why it matters
For continuous robot operation, the authors target geometric accuracy (trusted extrinsic calibration), appearance fidelity (global reconstruction benefits), and re-enterability (pose recovery after tracking loss). By quantifying how calibration quality propagates into reconstruction and relocalization, the method aims to make the 3DGS map dependable for localization, mapping, and downstream decision-making.
Method
- Estimate LiDAR–camera extrinsic offline using direct-method LiDAR–camera alignment on dedicated calibration segments; hold extrinsic fixed for online and offline tracks.
- Refine keyframe poses and Gaussian map alternately under a gated differentiable-rendering and LiDAR-projection objective, then rebuild global reconstruction from the refined keyframe schema with online constraints lifted.
- On tracking loss, recover pose from the final 3DGS map using an XFeat bypass map, RANSAC-PnP coarse localization, and Gaussian RGB-D warp refinement.
Limitation
The relocalization module can fail to recover accurate poses in extreme weak-texture scenes due to lack of sufficient features for matching and pose estimation (e.g., Bright achieves 0% success).
Abstract (from arXiv)
Photorealistic and metrically consistent world representations are a primary requirement for LiDAR--inertial--visual SLAM systems that serve robots over continuous operation. 3D Gaussian splatting makes such representations attainable by supplying an explicit, renderable, and queryable scene model. We present GEMOR-GS (Global Extrinsic-calibration-refined Mapping and Online Relocalization for 3D Gaussian-Splatting SLAM) to make the Gaussian map geometrically accurate, globally refined, and re-enterable, providing a better appearance for robot perception and decision-making. To improve the 3DGS map, the LiDAR--camera extrinsic is first estimated offline by direct-method alignment over dedicated calibration segments and reused by all following online and offline processes. Under the calibrated geometry, keyframe poses and Gaussian parameters are refined alternately under a gated differentiable-rendering and LiDAR--projection objective. The final global reconstruction is then rebuilt from the refined keyframe schema with all online constraints lifted. For the robot SLAM lifecycle, the method recovers the pose from the 3DGS map after tracking loss through an XFeat bypass map, RANSAC-PnP coarse localization, and Gaussian RGB-D warp refinement. Experiments on one self-collected and five public sequences quantify each element: the offline reconstruction wins all paired keyframes against the online map by 14.3--19.9 dB, improves PSNR by approximately 9 dB over Gaussian-LIC2, and achieves 86--100% held-out relocalization success tied to extrinsic quality. The results show that GEMOR-GS improves the geometric accuracy, appearance fidelity, and re-enterability of the 3DGS map, serving as a dependable robot world model for SLAM and downstream tasks.
Related papers
- Into the LidarVerse: A Real-Time Unified LiDAR World Model Across Tasks and Sensors
- LynxScenes: A Large-Scale Real-World Benchmark for Outdoor Wheel-Legged Robot Perception, Navigation, Imagination and Reconstruction
- G$^2$WAM: World Action Models for Efficient and Geometry Grounded Image-Goal Navigation
- UniFlowWAM: A Unified World-Action Model with Visual Semantics and 4D Geometry
- Let LiDAR Decide Where to Look: Geometry-Measured Token Selection for Efficient Sparse Latent World Models
Publication record
Bibliographic evidence describes where and when a record appeared. It does not validate scientific claims.
First public release: 2026-11-10 · Indexed by Pub-AI: 2026-11-11 UTC
GEMOR-GS: Global Extrinsic-calibration-refined Mapping and Online Relocalization for 3D Gaussian-Splatting SLAM
Preprint posted: 2026-11-10 · Original record ↗
- arxiv
- 2611.13528
Explicit source versions (1)
- v1 · 2026-11-10
- Preprint available
Source date: 2026-11-10 · First detected: 2026-11-11 UTC · Historical baseline
Source evidence (2)
- Official evidence ↗ · RSS guid=oai:arXiv.org:2611.13528v1; arxiv:announce_type=new; item pubDate (official metadata)
- Official evidence ↗ · atom:published (official metadata)
Abstract · arxiv
Photorealistic and metrically consistent world representations are a primary requirement for LiDAR--inertial--visual SLAM systems that serve robots over continuous operation. 3D Gaussian splatting makes such representations attainable by supplying an explicit, renderable, and queryable scene model. We present GEMOR-GS (Global Extrinsic-calibration-refined Mapping and Online Relocalization for 3D Gaussian-Splatting SLAM) to make the Gaussian map geometrically accurate, globally refined, and re-enterable, providing a better appearance for robot perception and decision-making. To improve the 3DGS map, the LiDAR--camera extrinsic is first estimated offline by direct-method alignment over dedicated calibration segments and reused by all following online and offline processes. Under the calibrated geometry, keyframe poses and Gaussian parameters are refined alternately under a gated differentiable-rendering and LiDAR--projection objective. The final global reconstruction is then rebuilt from the refined keyframe schema with all online constraints lifted. For the robot SLAM lifecycle, the method recovers the pose from the 3DGS map after tracking loss through an XFeat bypass map, RANSAC-PnP coarse localization, and Gaussian RGB-D warp refinement. Experiments on one self-collected and five public sequences quantify each element: the offline reconstruction wins all paired keyframes against the online map by 14.3--19.9 dB, improves PSNR by approximately 9 dB over Gaussian-LIC2, and achieves 86--100% held-out relocalization success tied to extrinsic quality. The results show that GEMOR-GS improves the geometric accuracy, appearance fidelity, and re-enterability of the 3DGS map, serving as a dependable robot world model for SLAM and downstream tasks.
