Skip to content

Latest commit

 

History

History
43 lines (27 loc) · 1.9 KB

README.md

File metadata and controls

43 lines (27 loc) · 1.9 KB

SLAM library

The SLAM library is designed based on ORB-SLAM3 released under GPLv3 license.

  • [ORB-SLAM3] Carlos Campos, Richard Elvira, Juan J. Gómez Rodríguez, José M. M. Montiel, and Juan D. Tardós, ORB-SLAM3: An Accurate Open-Source Library for Visual, Visual-Inertial and Multi-Map SLAM, IEEE Transactions on Robotics 37(6):1874-1890, Dec. 2021. PDF.

Prerequisites

We use Ubuntu 22.04. It would be easy to build this library on other Linux systems. If there is any problem, users may first try to build ORB-SLAM3 on the same system to get some experience.

Configuration

Open src/SLAM.cc, modify the following lines (L11~14):

bool use_viewer = false;     // whether to use ORB-SLAM3 viewer
string vocab_file = "<absolute path>/slam/Vocabulary/ORBvoc.txt";
string cam_settings_dir = "<absolute path>/slam/CameraSettings/";     // end with '/'
string map_points_save_dir = "<absolute output path>/";    // end with '/', output path

Note: you should use absolute paths. Set use_viewer=true to open the ORB-SLAM3 built-in visualization.

The camera parameters and some other SLAM options are set in the camera setting .yaml file.

Build

Make sure all dependencies have been installed and the paths have been correctly set. Then, execute:

cd slam
chmod +x build.sh
./build.sh

This will automatically build the Thirdparty libraries and finally create libslam.so in the lib/ folder. If you move the generated library file, edit wrapper.py to set the correct library file path slam_library_file.