/MF-SLAM

[MF-SLAM] Mingchi Feng, Jinglin Liu, Xin Wang, Chengnan Li

Primary LanguageC++GNU General Public License v3.0GPL-3.0

MF-SLAM

Beta version, 21 October 2020

Authors: Mingchi Feng, Jinglin Liu, Xin Wang, Chengnan Li, MF-SLAM: Multi-focal SLAM

MF-SLAM is a multi focal length visual slam, which can combine two cameras with different focal lengths to form stereo vision, or two cameras with the same focal length. It is improved on ORB-SLAM3, thank them very much for ORB-SLAM3's work and great efforts.

DOI: 10.1007/978-3-030-89134-3_45

We provide examples to run MF-SLAM in the KITTI and self-recorded dataset 提取码: j3w7

MF-SLAM is based on ORB-SLAM3 developed by Carlos Campos, Richard Elvira, Juan J. Gómez Rodríguez, José M. M. Montiel, Juan D. Tardos.

Video

China: billbill

World: youtube

KITTI dataset experiment

Self-recorded dataset experiment

6mm&12mm12mm&16mm16mm&25mm

playgroundAcademic_Buildinglibrary

1. License

MF-SLAM is released under GPLv3 license. For a list of all code/library dependencies (and associated licenses), please see Dependencies.md.

For a closed-source version of MF-SLAM for commercial purposes, please contact the authors: Mingchi Feng

If you use MF-SLAM in an academic work, please cite:

@article{ORBSLAM3_2020,
  title={{ORB-SLAM3}: An Accurate Open-Source Library for Visual, Visual-Inertial 
           and Multi-Map {SLAM}},
  author={Campos, Carlos AND Elvira, Richard AND G\´omez, Juan J. AND Montiel, 
          Jos\'e M. M. AND Tard\'os, Juan D.},
  journal={arXiv preprint arXiv:2007.11898},
  year={2020}
  
  title={{MF-SLAM}: Multi-focal SLAM {SLAM}},
  author={ Mingchi Feng, Jinglin Liu, Xin Wang, Chengnan Li.},
  conference={ICIRA 2021: Intelligent Robotics and Applications},
  year={2021}
 }

2. Prerequisites

We have tested the library in Ubuntu 16.04 , 18.04 and 18.04, but it should be easy to compile in other platforms. A powerful computer (e.g. i7) will ensure real-time performance and provide more stable and accurate results.

C++11 or C++0x Compiler

We use the new thread and chrono functionalities of C++11.

Pangolin

We use Pangolin for visualization and user interface. Dowload and install instructions can be found at: https://github.com/stevenlovegrove/Pangolin.

OpenCV

We use OpenCV to manipulate images and features. Dowload and install instructions can be found at: http://opencv.org. Required at leat 3.0. Tested with OpenCV 3.2.0.

Eigen3

Required by g2o (see below). Download and install instructions can be found at: http://eigen.tuxfamily.org. Required at least 3.1.0.

DBoW2 and g2o (Included in Thirdparty folder)

We use modified versions of the DBoW2 library to perform place recognition and g2o library to perform non-linear optimizations. Both modified libraries (which are BSD) are included in the Thirdparty folder.

Python

Required to calculate the alignment of the trajectory with the ground truth. Required Numpy module.

3. Building MF-SLAM library and examples

Clone the repository:

git clone https://github.com/JinLn/MF-SLAM.git MF-SLAM

We provide a script build.sh to build the Thirdparty libraries and MF-SLAM. Please make sure you have installed all required dependencies (see section 2). Execute:

cd MF-SLAM
chmod +x build.sh
./build.sh

This will create libMF-SLAM.so at lib folder and the executables in Examples folder.

4. KITTI Examples

KITTI dataset was provide 11 sequences (00-10) with ground truth trajectories for training and 11 sequences (11-21) without ground truth for evaluation. For this benchmark you may provide results using monocular or stereo visual odometry, laser-based SLAM or algorithms that combine visual and LIDAR information. We provide an example script to launch KITTI sequences in all the sensor configurations.

  1. Download a sequence from http://www.cvlibs.net/datasets/kitti/eval_odometry.php
   ./stereo_test ../../Vocabulary/ORBvoc.txt /home/jinln/jinln/DATASET/kitti/MF-dyna_yaml/KITTI_stereo_00-02.yaml /home/jinln/jinln/DATASET/kitti/00
  1. Evaluation using evo tools

Evaluation

 evo_traj kitti KITTI_00_ORB.txt MF-SLAM_00.txt --ref=KITTI_00_gt.txt -p --plot_mode=xz

5. Self-recorded dataset Examples

Use two different focal length lenses to form stereo vision (recommended focal length ratio not to exceed 2 times), use multi-focal-stereo-calib to calibrate binocular camera, and the obtained parameters replace the parameters of the .yaml file.

    ./stereo_test ../../Vocabulary/ORBvoc.txt /home/jinln/jinln/DATASET/baslerstereodata/20210414/12-16/MF_01.yaml /home/jinln/jinln/DATASET/baslerstereodata/20210414/12-16/2021041404