Multicam-SLAM: Non-overlapping Multi-camera SLAM for Indirect Visual Localization and Navigation

TL;DR

Multicam-SLAM addresses the limitations of single-camera visual SLAM by deploying a non-overlapping multicamera RGB-D setup that provides richer spatial information. It introduces a full MKF-based SLAM pipeline with a novel multicam entity, a unified map (comprising 3D points, MKFs, and a covisibility graph), and an on-the-fly extrinsic calibration via pose-graph optimization. The approach demonstrates improved accuracy and robustness over a strong RGB-D baseline (ORB-SLAM2) in both mobile and handheld experiments, supported by detailed runtime and reliability analyses. This work advances visual localization and navigation in challenging environments by enabling scalable multi-camera fusion with end-to-end SLAM, including loop closure and backend optimization.

Abstract

This paper presents a novel approach to visual simultaneous localization and mapping (SLAM) using multiple RGB-D cameras. The proposed method, Multicam-SLAM, significantly enhances the robustness and accuracy of SLAM systems by capturing more comprehensive spatial information from various perspectives. This method enables the accurate determination of pose relationships among multiple cameras without the need for overlapping fields of view. The proposed Muticam-SLAM includes a unique multi-camera model, a multi-keyframes structure, and several parallel SLAM threads. The multi-camera model allows for the integration of data from multiple cameras, while the multi-keyframes and parallel SLAM threads ensure efficient and accurate pose estimation and mapping. Extensive experiments in various environments demonstrate the superior accuracy and robustness of the proposed method compared to conventional single-camera SLAM systems. The results highlight the potential of the proposed Multicam-SLAM for more complex and challenging applications. Code is available at \url{https://github.com/AlterPang/Multi_ORB_SLAM}.

Figures (8)

• Figure 1: Overview of the Proposed Multicam-SLAM system. The system is structured into three concurrent threads: pose tracking, local mapping, and loop closure detection.
• Figure 2: Illustration of the Multiple Camera Coordinate System.
• Figure 3: Pose Graph Optimization for multicam entity. (a) The initial pose graph. The black and brown vertices represent the poses of the two cameras respectively. The blue and green edges are the transformations between adjacent key frames of the two cameras respectively. (b) Matching-based Edge Establishment. The red edges represent The matching keyframes, the purple edge represents the final extrinsic parameters between the initial keyframes of the two cameras.
• Figure 4: Illustration of the Multiple Camera Hardware.
• Figure 5: Illustration of the Mobile Platform for Evaluation.