Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

BundledSLAM: An Accurate Visual SLAM System Using Multiple Cameras

Han Song, Cong Liu, Huafeng Dai

TL;DR

BundledSLAM extends ORB-SLAM2 to multi-camera visual SLAM by mapping measurements from all cameras into a virtual BundledFrame and performing bundle adjustment with fixed inter-camera transforms. The approach introduces BundledFrame, BundledKeyframe, and BundledMap to unify multi-camera data, and employs motion-only, local, and global bundle adjustment alongside a loop-closure pipeline based on DBoW2 to improve pose and map accuracy. Evaluations on the EuRoC dataset show BundledSLAM achieving consistently higher accuracy and lower APE than ORB-SLAM2 and VINS-Stereo, validating the benefits of cross-camera data fusion for feature matching, map reuse, and loop closure. The work highlights the potential of multi-camera fusion to enhance SLAM robustness, with future directions including IMU fusion and strategies to mitigate computational overhead while preserving performance.

Abstract

Multi-camera SLAM systems offer a plethora of advantages, primarily stemming from their capacity to amalgamate information from a broader field of view, thereby resulting in heightened robustness and improved localization accuracy. In this research, we present a significant extension and refinement of the state-of-the-art stereo SLAM system, known as ORB-SLAM2, with the objective of attaining even higher precision. To accomplish this objective, we commence by mapping measurements from all cameras onto a virtual camera termed BundledFrame. This virtual camera is meticulously engineered to seamlessly adapt to multi-camera configurations, facilitating the effective fusion of data captured from multiple cameras. Additionally, we harness extrinsic parameters in the bundle adjustment (BA) process to achieve precise trajectory estimation.Furthermore, we conduct an extensive analysis of the role of bundle adjustment (BA) in the context of multi-camera scenarios, delving into its impact on tracking, local mapping, and global optimization. Our experimental evaluation entails comprehensive comparisons between ground truth data and the state-of-the-art SLAM system. To rigorously assess the system's performance, we utilize the EuRoC datasets. The consistent results of our evaluations demonstrate the superior accuracy of our system in comparison to existing approaches.

BundledSLAM: An Accurate Visual SLAM System Using Multiple Cameras

TL;DR

BundledSLAM extends ORB-SLAM2 to multi-camera visual SLAM by mapping measurements from all cameras into a virtual BundledFrame and performing bundle adjustment with fixed inter-camera transforms. The approach introduces BundledFrame, BundledKeyframe, and BundledMap to unify multi-camera data, and employs motion-only, local, and global bundle adjustment alongside a loop-closure pipeline based on DBoW2 to improve pose and map accuracy. Evaluations on the EuRoC dataset show BundledSLAM achieving consistently higher accuracy and lower APE than ORB-SLAM2 and VINS-Stereo, validating the benefits of cross-camera data fusion for feature matching, map reuse, and loop closure. The work highlights the potential of multi-camera fusion to enhance SLAM robustness, with future directions including IMU fusion and strategies to mitigate computational overhead while preserving performance.

Abstract

Multi-camera SLAM systems offer a plethora of advantages, primarily stemming from their capacity to amalgamate information from a broader field of view, thereby resulting in heightened robustness and improved localization accuracy. In this research, we present a significant extension and refinement of the state-of-the-art stereo SLAM system, known as ORB-SLAM2, with the objective of attaining even higher precision. To accomplish this objective, we commence by mapping measurements from all cameras onto a virtual camera termed BundledFrame. This virtual camera is meticulously engineered to seamlessly adapt to multi-camera configurations, facilitating the effective fusion of data captured from multiple cameras. Additionally, we harness extrinsic parameters in the bundle adjustment (BA) process to achieve precise trajectory estimation.Furthermore, we conduct an extensive analysis of the role of bundle adjustment (BA) in the context of multi-camera scenarios, delving into its impact on tracking, local mapping, and global optimization. Our experimental evaluation entails comprehensive comparisons between ground truth data and the state-of-the-art SLAM system. To rigorously assess the system's performance, we utilize the EuRoC datasets. The consistent results of our evaluations demonstrate the superior accuracy of our system in comparison to existing approaches.
Paper Structure (11 sections, 10 equations, 6 figures, 1 table)

This paper contains 11 sections, 10 equations, 6 figures, 1 table.

Table of Contents

  1. Introduction
  2. BundledSLAM
  3. System Overview
  4. BundledFrame, BundledKeyframe and BundledMap
  5. Camera Projection Model with Multiple Cameras
  6. Motion Estimation and Bundle Adjustment with Multiple Cameras
  7. Loop Closing
  8. Evaluation
  9. EuRoC Datasets
  10. Accuracy Evaluation
  11. Conclusion

Figures (6)

  • Figure 1: Pipeline of BundledSLAM
  • Figure 2: An example of Bundled, comprising unique feature IDs, monocular points, and matched points. Number of Unique Feature = Number of Monocular Feature + Number of Matched Feature
  • Figure 3: General Overview of BundledFrame
  • Figure 4: Visual odometry with multiple synchronized cameras. The goal is to estimate the relative motion of the first camera $\bm {C}_1^k$ in BundledFrame or BundledKeyframe at each moment to world coordinates.
  • Figure 5: Trajectory in EuRoC dataset compared with ORB-SLAM2 and VINS-Stereo.
  • ...and 1 more figures