This repository contains a proof of concept for a SVO based stereo camera SLAM library.
The repository is organized as follows:
| Direcotry | Description |
|---|---|
| doc | Documentation of the algorithm, implementation, etc. |
| src | Source code of library, test application, demo, qt viewer and python wrapper |
| test | Test scripts, test videos, etc. |
The source code includes some doxygen comments. Check out the documentation folder for indepth information.
The thesis and source code documentation can be found here:
The library allows to process stereo images and calculates the camera position based on this images.
The test application allows to process input images from different sources like Econ Tara Camera, EuRoC dataset or video input. It requires a YAML file with camera parameters. See src/app/Blender.yaml for more details.
The demo application is a simple ar-application which shows what a SLAM library can do. It only supports Econ Tara an requires a YAML file with camera settings (src/app/Econ.yaml).
The Qt 3D viewer can connect to the test application to show keyframes, current pose and trajectory.