MGS-SLAM: Monocular Sparse Tracking and Gaussian Mapping with Depth Smooth Regularization

TL;DR

This paper presents MGS-SLAM, a monocular SLAM framework that jointly optimizes visual odometry and Gaussian map reconstruction, significantly improving accuracy and fidelity in pose estimation and scene understanding.

Contribution

It introduces a novel joint optimization approach with depth smoothness and SDAR to enhance Gaussian Splatting in monocular SLAM, addressing geometric accuracy and tracking limitations.

Findings

Achieves state-of-the-art pose estimation accuracy.

Outperforms previous monocular methods in view synthesis.

Improves geometric reconstruction fidelity.

Abstract

This letter introduces a novel framework for dense Visual Simultaneous Localization and Mapping (VSLAM) based on Gaussian Splatting. Recently, SLAM based on Gaussian Splatting has shown promising results. However, in monocular scenarios, the Gaussian maps reconstructed lack geometric accuracy and exhibit weaker tracking capability. To address these limitations, we jointly optimize sparse visual odometry tracking and 3D Gaussian Splatting scene representation for the first time. We obtain depth maps on visual odometry keyframe windows using a fast Multi-View Stereo (MVS) network for the geometric supervision of Gaussian maps. Furthermore, we propose a depth smooth loss and Sparse-Dense Adjustment Ring (SDAR) to reduce the negative effect of estimated depth maps and preserve the consistency in scale between the visual odometry and Gaussian maps. We have evaluated our system across various synthetic and real-world datasets. The accuracy of our pose estimation surpasses existing methods and achieves state-of-the-art. Additionally, it outperforms previous monocular methods in terms of novel view synthesis and geometric reconstruction fidelities.