Globally Consistent RGB-D SLAM with 2D Gaussian Splatting

TL;DR

This paper introduces 2DGS-SLAM, a novel RGB-D SLAM system that uses 2D Gaussian splatting to improve geometric accuracy, consistency, and efficiency in 3D reconstruction and mapping.

Contribution

The paper proposes a new SLAM system leveraging 2D Gaussian splatting for better depth consistency, loop closure, and real-time global mapping.

Findings

Achieves superior tracking accuracy compared to existing methods.

Provides higher surface reconstruction quality.

Maintains high-fidelity image rendering with improved efficiency.

Abstract

Recently, 3D Gaussian splatting-based RGB-D SLAM displays remarkable performance of high-fidelity 3D reconstruction. However, the lack of depth rendering consistency and efficient loop closure limits the quality of its geometric reconstructions and its ability to perform globally consistent mapping online. In this paper, we present 2DGS-SLAM, an RGB-D SLAM system using 2D Gaussian splatting as the map representation. By leveraging the depth-consistent rendering property of the 2D variant, we propose an accurate camera pose optimization method and achieve geometrically accurate 3D reconstruction. In addition, we implement efficient loop detection and camera relocalization by leveraging MASt3R, a 3D foundation model, and achieve efficient map updates by maintaining a local active map. Experiments show that our 2DGS-SLAM approach achieves superior tracking accuracy, higher surface reconstruction quality, and more consistent global map reconstruction compared to existing rendering-based SLAM methods, while maintaining high-fidelity image rendering and improved computational efficiency.