GLAM-SLAM pairs ORB-SLAM2 with Gaussian mapping for real-time large-scale reconstruction
zhenjun_zhao · x · 2026-07-24
GLAM-SLAM is a real-time Gaussian large-scale mapping system built on a decoupled SLAM pipeline.
The method combines:
- an ORB-SLAM2 thread for real-time tracking and local mapping,
- a Gaussian Mapper thread that builds the anchor grid from sparse ORB-SLAM2 points plus flow-densified points,
- and a 3D Gaussian optimizer using photometric loss.
The paper reports reconstruction-quality results on KITTI, Oxford, and Malaga datasets, and compares against PhotoSLAM, GigaSLAM, and MonoGS. The attached tables suggest the method is aiming at practical, scalable visual mapping rather than a toy demo.
More from Embodied
- SAGE generates simulation-ready 3D scenes and releases a 10k embodied-AI dataset — rsasaki0109 · 2026-07-24
- ReferTrack tracks language-specified targets with one camera and reaches 89.4% on EVT-Bench — tencent · 2026-07-24
- IGGT4D brings streaming 4D instance-grounded geometry to dynamic scene understanding — zhenjun_zhao · 2026-07-24
- MIT report: nuclear plants are shifting toward supervised autonomous control — nordicinst · 2026-07-24
- Black Forest Labs video models are being used to automate Audi manufacturing — hsu_byron · 2026-07-24
- Hong Kong prepares for its first test run of fully driverless vehicles — Baidu_Inc · 2026-07-24