Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

XRDSLAM: A Flexible and Modular Framework for Deep Learning based SLAM

Xiaomeng Wang, Nan Wang, Guofeng Zhang

TL;DR

A flexible SLAM framework that integrates several state-of-the-art SLAM algorithms with different types, including NeRF and 3DGS based SLAM, and even odometry or reconstruction algorithms, which demonstrates the flexibility and extensibility.

Abstract

In this paper, we propose a flexible SLAM framework, XRDSLAM. It adopts a modular code design and a multi-process running mechanism, providing highly reusable foundational modules such as unified dataset management, 3d visualization, algorithm configuration, and metrics evaluation. It can help developers quickly build a complete SLAM system, flexibly combine different algorithm modules, and conduct standardized benchmarking for accuracy and efficiency comparison. Within this framework, we integrate several state-of-the-art SLAM algorithms with different types, including NeRF and 3DGS based SLAM, and even odometry or reconstruction algorithms, which demonstrates the flexibility and extensibility. We also conduct a comprehensive comparison and evaluation of these integrated algorithms, analyzing the characteristics of each. Finally, we contribute all the code, configuration and data to the open-source community, which aims to promote the widespread research and development of SLAM technology within the open-source ecosystem.

XRDSLAM: A Flexible and Modular Framework for Deep Learning based SLAM

TL;DR

A flexible SLAM framework that integrates several state-of-the-art SLAM algorithms with different types, including NeRF and 3DGS based SLAM, and even odometry or reconstruction algorithms, which demonstrates the flexibility and extensibility.

Abstract

In this paper, we propose a flexible SLAM framework, XRDSLAM. It adopts a modular code design and a multi-process running mechanism, providing highly reusable foundational modules such as unified dataset management, 3d visualization, algorithm configuration, and metrics evaluation. It can help developers quickly build a complete SLAM system, flexibly combine different algorithm modules, and conduct standardized benchmarking for accuracy and efficiency comparison. Within this framework, we integrate several state-of-the-art SLAM algorithms with different types, including NeRF and 3DGS based SLAM, and even odometry or reconstruction algorithms, which demonstrates the flexibility and extensibility. We also conduct a comprehensive comparison and evaluation of these integrated algorithms, analyzing the characteristics of each. Finally, we contribute all the code, configuration and data to the open-source community, which aims to promote the widespread research and development of SLAM technology within the open-source ecosystem.

Paper Structure

This paper contains 37 sections, 7 figures, 2 tables.

Table of Contents

  1. INTRODUCTION
  2. Related Work
  3. Learning based SLAM
  4. Frameworks of SLAM
  5. System
  6. Framework Overview
  7. Dataset Input
  8. Core SLAM Algorithms
  9. Visualization
  10. Evaluation
  11. Multipie Processes
  12. Tracker
  13. Mapper
  14. Visualizer
  15. Data sharing mechanism
  16. ...and 22 more sections

Figures (7)

  • Figure 1: System Overview of XRDSLAM. It takes RGB, depth, IMU, and other data as Input, processed through the Tracker and Mapper processes, with core SLAM algorithm modules providing essential functionality to these processes. A reusable Visualizer module is used to display the algorithm's outputs, and an Evaluation module is included for comprehensive metric assessment. The section within the dashed box needs to be customized by developers.
  • Figure 2: Components of XRDSLAM.
  • Figure : NICE-SLAM
  • Figure : NICE-SLAM
  • Figure : Co-SLAM
  • ...and 2 more figures