GeoGS-SLAM is presented, an online monocular dense reconstruction system that combines the 3DGS-based map representation with learned geometric priors and achieves superior rendering quality and tracking accuracy compared to state-of-the-art methods while maintaining online real-time performance.
Abstract
SLAM methods based on 3D Gaussian Splatting (3DGS) have demonstrated impressive tracking and mapping performance, but typically require additional geometric information from external depth sensors. Meanwhile, recent SLAM systems that leverage geometric priors from pre-trained feed-forward models enable real-time dense reconstruction, yet often discard original RGB information during optimization, thus degrading overall reconstruction quality. We present GeoGS-SLAM, an online monocular dense reconstruction system that combines the 3DGS-based map representation with learned geometric priors. Given uncalibrated RGB input, we first employ a feed-forward visual geometry model to predict camera and scene priors. The Gaussian scene map is then expanded by directly sampling Gaussian primitives from both RGB input and geometric priors. Camera poses and the scene map are jointly optimized through a coarse-to-fine strategy that minimizes both photometric and geometric losses. To ensure global consistency, we further incorporate online loop closure detection and pose graph optimization. Extensive experiments across indoor and outdoor benchmarks demonstrate that GeoGS-SLAM achieves superior rendering quality and tracking accuracy compared to state-of-the-art methods while maintaining online real-time performance. Project page: https://rlgao.github.io/geogs_slam.
Existing neural SLAM and 3D Gaussian Splatting (3DGS) SLAM systems often suffer from insufficient observations during indoor turn-arounds and large viewpoint changes, leading to incomplete coverage and missing details in corner regions. We propose IMGS-SLAM, a monocular Gaussian SLAM system tailored for indoor reconstruction, which improves mapping completeness and visual detail fidelity using only RGB images while maintaining competitive tracking accuracy. The method leverages a learning-based dense SLAM frontend to provide camera poses and dense geometric priors, and adopts 3DGS as the map representation. We propose a coverage-aware Dual-Cue mapping-frame selection strategy that decouples tracking keyframes from mapping frames and selects views with high expected coverage gain. This design improves map completeness under indoor turn-around motions and large viewpoint changes, while online-to-offline refinement further improves visual consistency and local details. In addition, we introduce quadtree-guided structured initialization and a high-frequency weighted loss to enhance textures and edge details, and incorporate co-visibility-constrained densification and pruning to reduce artifacts. Experiments on Replica, TUM RGB-D, and ScanNet demonstrate improved map completeness and competitive rendering quality in both synthetic and real indoor scenes.
Yaoming Zhuang, Yifan Chao, Tongrui Liu et al.· IEEE Robotics and Automation...· 0 citations
Dense visual SLAM is a fundamental problem in robotics. Recent advances in 3DGS have demonstrated its potential for dense SLAM. Existing 3DGS frameworks focus on both appearance and geometry modeling. However, scene geometry is typically more critical for SLAM than novel view synthesis because downstream robotic tasks, such as navigation and obstacle avoidance, rely primarily on accurate spatial geometry rather than photorealistic rendering. This observation raises a natural question: Is it feasible for 3DGS to perform 3D reconstruction without scene appearance modeling? Motivated by this, we propose Geometry-only Gaussian Splatting (GeoGS), which directly reconstructs scene geometry, and further present GeoGS-SLAM, a dense visual SLAM system built upon this representation. Specifically, GeoGS retains only spatial parameters to reduce the number of per-primitive parameters by over 80%. In contrast to existing 3DGS methods, GeoGS focuses solely on geometric reconstruction, which significantly reduces the number of Gaussian primitives, accelerates geometric convergence, and enhances robustness to illumination variations. In addition, we present an effective training framework that optimizes the Gaussian primitives via single-view and multi-view geometric and photometric supervision, and speeds up geometry convergence with a local-plane driven initialization that better aligns primitives with local structures. Furthermore, we introduce a map update strategy for loop closure that globally transforms the Gaussian map to align it with the corrected pose estimates, thereby preventing map tearing caused by inconsistent per-viewpoint pose corrections in existing methods. Extensive experiments on synthetic and real-world benchmarks demonstrate that our method outperforms SOTA methods in terms of online mapping efficiency and geometric reconstruction quality.
Lipu Zhou, Yaoyun Kang, Junxiang Pang et al.· 0 citations
3D Gaussian Splatting (3DGS) has gained prominence in autonomous driving and robotics for its rendering efficiency and high-fidelity reconstruction capabilities. However, incremental 3DGS map construction at large scales remains challenging due to sensor sparsity and computational constraints. In this paper, we propose LV-GS SLAM, a novel system that integrates LiDAR and visual data for incremental, large-scale reconstruction with real-time tracking. This system provides fast and robust pose estimation while also enabling photorealistic rendering. We first employ a LiDAR odometry frontend that processes 30 Hz LiDAR inputs to provide robust initial poses. In our implementation, the complete LiDAR tracking pipeline runs at 15–19 Hz, while the mapping module performs incremental optimization on selected keyframes. To address the sparsity-induced surface discontinuity in conventional LiDAR-based reconstruction, we propose a novel depth propagation approach that initializes 3D Gaussian primitives using dense depth maps, achieving faster PSNR convergence with 9× fewer optimization iterations compared to direct LiDAR initialization. Furthermore, we develop a keyframe-based submap management framework that dynamically adjusts memory allocation based on both primitive density and inter-frame overlap ratio, effectively preventing GPU memory overflow. Our system has been validated on the KITTI dataset, achieving superior rendering quality compared with representative reproducible baselines. We further validate the robustness of the system on a quadruped robot platform, demonstrating satisfactory performance in both pose estimation and high-fidelity reconstruction.
Recent 3D Gaussian Splatting (3DGS) has enabled efficient photorealistic view synthesis and is rapidly being adopted in simultaneous localization and mapping (SLAM) systems for online mapping. In these systems, a Gaussian map must be expanded and refined incrementally while tracking runs in real time, so initialization and density control directly determine where limited computation and iterations are spent. This contrasts with offline 3DGS reconstruction, where such heuristics can be amortized over long optimization schedules. However, most 3DGS-SLAM pipelines inherit initialization and density-control heuristics from offline reconstruction, which can become brittle under the strict per-keyframe optimization budgets and incremental map growth of online SLAM. In this work, we revisit these heuristics in a decoupled 3DGS-SLAM setting and propose three geometry-aware methods that operate in the mapping thread: transmittance-preserving densification, camera-aware scale initialization from depth and intrinsics, and error-guided densification that focuses new primitives on high-residual regions. Our results show consistent improvements in rendering quality with negligible overhead, highlighting the coupling between photometric residuals and pose uncertainty in online SLAM. We will open-source our code to the community to foster growth and validate reproducibility.
Thai Luu, Quan Tran, Hieu Phan et al.· 0 citations
Monocular RGB dense SLAM remains challenging because depth and global metric scale are not directly observable from a single camera. Existing systems often combine optical-flow and monocular-geometry priors predicted by independently trained networks, which can provide inconsistent constraints to bundle adjustment (BA). Our quantitative prior-consistency analysis indicates that this disagreement is an important contributor to unstable local optimization and reconstruction error rather than the sole cause of drift. We propose KP-SLAM, which predicts dense optical flow and paired pointmap priors from a shared representation and incorporates them into the same BA backend. We further introduce a Depth-Scale-Pose-to-Pointmap (DSPP) objective that relates optimized inverse depth, edge-wise relative scale, and camera pose to paired pointmap constraints. Experiments on ScanNet, TUM-RGBD, KITTI, Tanks-and-Temples, and dynamic sequences show improved tracking, depth, and rendering metrics over the compared RGB-only baselines under the reported settings. The results support the usefulness of synchronized priors while also revealing remaining limitations in highly dynamic, weakly textured, and large-scale scenes.
Song Gao, Xinyu Huang, Zheng Huang et al.· Symmetry· 0 citations
Long-sequence 3D reconstruction from RGB videos requires both accurate local geometry and globally consistent camera motion. Feed-forward models provide strong depth and pose predictions, but their memory cost prevents joint inference over long sequences. Chunk-wise processing improves scalability, yet independently predicted chunks often exhibit scale drift, pose errors, and point-cloud misalignment. We present GeoWeaver, a unified framework comprising a Geometric Prior Model (GPM) and Test-Time Adaptation (TTA). The GPM predicts chunk-wise depth, confidence, and camera parameters as adjustable geometric priors. TTA then performs sequential initialization, global chunk-level Sim(3) alignment, and coarse-to-fine refinement of camera poses, affine depth corrections, and intrinsics. Dense correspondences provide adjacent, cross-chunk, and long-range constraints, while a robust CDF-style objective jointly optimizes weighted 2D reprojection and 3D consistency residuals. This design preserves local geometric accuracy while correcting accumulated pose, scale, depth, and calibration errors. Experiments across diverse long-sequence benchmarks demonstrate improved camera accuracy, global consistency, and point-cloud quality. Ablations verify the contribution of each adaptation stage, and applying the same TTA procedure to different geometric prior models consistently improves their trajectory estimates, demonstrating that GeoWeaver is not tied to a specific GPM.
Tinghao Jiang, Sheng Tang, Shengzhe Wei et al.· 0 citations
