LDSO: Direct Sparse Odometry with Loop Closure

Related Publications

LDSO: Direct Sparse Odometry with Loop Closure, X. Gao, R. Wang, N. Demmel, D. Cremers, In International Conference on Intelligent Robots and Systems (IROS), 2018.
Direct Sparse Odometry, J. Engel, V. Koltun, D. Cremers, In IEEE Transactions on Pattern Analysis and Machine Intelligence (PAMI), 2018
A Photometrically Calibrated Benchmark For Monocular Visual Odometry, J. Engel, V. Usenko, D. Cremers, In arXiv:1607.02555, 2016

Dependencies

System dependencies

There is a convenience script that will help you install the needed libraries in Ubuntu 16.04 and later, including Eigen, glog, gtest, Suitesparse, OpenCV, libzip.

./install_dependencies.sh

On OSX you can install these via Homebrew.

Other libraries

Compile and install Pangolin for visualization.

Compile

./make_project.sh

This will build the thirdparty library and also ldso library for you. You can also follow the steps in this script manually (will compile DBoW3 and g2o first, and the ldso).

Run

We provide examples on three datasets.

TUM-Mono: https://vision.in.tum.de/mono-dataset
Kitti: Kitti odometry
EuRoC: Euroc MAV dataset

You can easily extend them in the examples folder or add your own executables for your camera or dataset.

After compilation, in the bin directory there will be three executables like run_dso_xxx. You can either specify the directories in the source file like examples/run_dso_xxx.cc, or pass them as command line parameters. When running LDSO, you will see a window showing an visualization of camera trajectory and tracked points.

Make sure your working directory is at the root of LDSO code, to ensure that files like the BoW vocabulary file are found.

TUM-Mono:

To run LDSO on TUM-Mono dataset sequence 34, execute:

./bin/run_dso_tum_mono \
    preset=0 \
    files=XXXXX/TUMmono/sequences/sequence_34/images.zip \
    vignette=XXXXX/TUMmono/sequences/sequence_34/vignette.png \
    calib=XXXXX/TUMmono/sequences/sequence_34/camera.txt \
    gamma=XXXXX/TUMmono/sequences/sequence_34/pcalib.txt

Kitti:

To run LDSO on Kitti dataset sequence 00, execute:

./bin/run_dso_kitti \
    preset=0 \
    files=XXXXX/Kitti/odometry/dataset/sequences/00/ \
    calib=./examples/Kitti/Kitti00-02.txt

EuRoC:

To run LDSO on EuRoC dataset sequence MH_01_easy, execute:

./bin/run_dso_euroc \
    preset=0 \
    files=XXXX/EuRoC/MH_01_easy/mav0/cam0/

Notes

LDSO is a monocular VO based on DSO with Sim(3) loop closing function. Note we still cannot know the real scale of mono-slam. We only make it more consistent in long trajectories.
The red line in pangolin windows shows the trajectory before loop closing, and the yellow line shows the trajectory after optimization.
If you are looking for code instructions, take a look at doc/notes_on_ldso.pdf and see if it can help you.
Set setting_enableLoopClosing to true/false to turn on/off loop closing function. setting_fastLoopClosing will record less data and make the loop closing faster.
If you need loop closing, please set setting_pointSelection=1 to make the program compute feature descriptors. If setting_pointSelection=0, the program acts just like DSO, and setting_pointSelection=2 means random point selection, which is faster but unstable.
Some of the GUI buttons may not work.
You might also want to have a look DSO's README: https://github.com/JakobEngel/dso/blob/master/README.md

License

LDSO, like DSO, is licensed under GPLv3. It makes use of several third-party libraries. Among other's it comes with the source of:

Sophus (MIT)
g2o (BSD, GPL, LGPL, ...)
DBoW3 (BSD)

tum-vision / LDSO