For more information see https://vision.in.tum.de/research/vslam/basalt
This project contains tools for:
- Camera, IMU and motion capture calibration.
- Visual-inertial odometry and mapping.
- Simulated environment to test different components of the system.
Some reusable components of the system are available as a separate header-only library (Documentation).
There is also a Github mirror of this project to enable easy forking.
Visual-Inertial Odometry and Mapping:
- Visual-Inertial Mapping with Non-Linear Factor Recovery, V. Usenko, N. Demmel, D. Schubert, J. Stückler, D. Cremers, In IEEE Robotics and Automation Letters (RA-L) [DOI:10.1109/LRA.2019.2961227] [arXiv:1904.06504].
Calibration (explains implemented camera models):
- The Double Sphere Camera Model, V. Usenko and N. Demmel and D. Cremers, In 2018 International Conference on 3D Vision (3DV), [DOI:10.1109/3DV.2018.00069], [arXiv:1807.08957].
Calibration (demonstrates how these tools can be used for dataset calibration):
- The TUM VI Benchmark for Evaluating Visual-Inertial Odometry, D. Schubert, T. Goll, N. Demmel, V. Usenko, J. Stückler, D. Cremers, In 2018 International Conference on Intelligent Robots and Systems (IROS), [DOI:10.1109/IROS.2018.8593419], [arXiv:1804.06120].
Calibration (describes B-spline trajectory representation used in camera-IMU calibration):
- Efficient Derivative Computation for Cumulative B-Splines on Lie Groups, C. Sommer, V. Usenko, D. Schubert, N. Demmel, D. Cremers, In 2020 Conference on Computer Vision and Pattern Recognition (CVPR), [DOI:10.1109/CVPR42600.2020.01116], [arXiv:1911.08860].
Optimization (describes square-root optimization and marginalization used in VIO/VO):
- Square Root Marginalization for Sliding-Window Bundle Adjustment, N. Demmel, D. Schubert, C. Sommer, D. Cremers, V. Usenko, In 2021 International Conference on Computer Vision (ICCV), [arXiv:2109.02182]
Install the latest published release into ~/.local:
curl -LsSf https://gitlab.com/VladyslavUsenko/basalt/-/raw/master/scripts/install.sh | sh
The installer places binaries in ~/.local/bin, libraries in ~/.local/lib, and data files in ~/.local/etc/basalt.
Clone the source code with the thirdparty/vcpkg submodule, then build with CMake presets. Install CMake (>= 3.24), Ninja, and a C++ compiler first.
git clone --recursive https://gitlab.com/VladyslavUsenko/basalt.git
cd basalt
# If you cloned without --recursive, fetch the vcpkg submodule:
# git submodule update --init thirdparty/vcpkg
# Bootstrap vcpkg once if needed:
# ./thirdparty/vcpkg/bootstrap-vcpkg.sh -disableMetrics
cmake --preset relwithdebinfo
cmake --build --preset relwithdebinfo -j8
ctest --preset relwithdebinfo
By default presets use:
thirdparty/vcpkg/scripts/buildsystems/vcpkg.cmake
On macOS, build the release preset and then package the release artifact:
cmake --preset release
cmake --build --preset release -j8
./scripts/package_macos_release.sh <tag>
- Camera, IMU and Mocap calibration. (TUM-VI, Euroc, UZH-FPV and Kalibr datasets)
- Visual-inertial odometry and mapping. (TUM-VI and Euroc datasets)
- Visual odometry (no IMU). (KITTI dataset)
- Simulation tools to test different components of the system.
- Batch evaluation tutorial (ICCV'21 experiments)
The code is provided under a BSD 3-clause license. See the LICENSE file for details. Note also the different licenses of thirdparty code.
Some improvements are ported back from the fork granite (MIT license).