This is YGZ-stereo-inertial SLAM, a stereo inertial VO code. It is designed for stereo and stereo-inertial sensor modules like vi-sensor. It uses a LK optical flow as front-end and a sliding window bundle adjustment as a backend. Feel free to try it in datasets and your own sensors.

The code is experimental and I can't guarantee its performance in all cases.


If you are using ubuntu, just type "./install_dep.sh" to install all the dependencies except pangolin.


run "./generate.sh" to compile all the things, or follow the steps in generate.sh


You can put stereo or stereo-imu data into ygz-stereo, for example the EUROC dataset (http://projects.asl.ethz.ch/datasets/doku.php?id=kmavvisualinertialdatasets). We provide the pure stereo vision and stereo-inertial vision for EUROC. Run the pure vision examples by typing:

bin/EurocStereo ./examples/EurocStereo.yaml

to run the pure vision mode. Don't forget to specify the dataset directory in the yaml config file first. Also, to run visual-inertial mode, type:

bin/EurocStereoVIO ./examples/EurocStereoVIO.yaml

to run the stereo VIO case.

Other things

YGZ-stereo is more robust than the previous YGZ-ORB-SLAM (and the code is also more clear). In EUROC it can pass the test of all MHxx and V101, V201, V202. For difficult cases it may still fail. As I will go to TUM soon, my future work may be a visual-inertial DSO.

YGZ stands for Yi-Guo-Zhou (a port of porridge, a group of mess) because it contains feature method, direct method and imu things.

The Note.md is a file of develop records.

Contact me ([email protected]) or Wang Jing (https://github.com/jingpang) for commercial use.

Thanks the following companies/people for financial support: