LSD-SLAM is a novel approach to real-time monocular SLAM. It is fully direct (i.e. does not use keypoints / features) and creates large-scale, semi-dense maps in real-time on a laptop. For more information see http://vision.in.tum.de/lsdslam where you can also find the corresponding publications and Youtube videos, as well as some example-input datasets, and the generated output as rosbag or .ply point cloud.
The source code is tested on Ubuntu 16.04 using cmake (3.5.1) and it depends on the following libraries:
- OpenCV 2.4
- g20 (commit 9b41a4e on Apr 6) for debugging purpose I've built g2o using the -DCMAKE_BUILD_TYPE=Debug
- Eigen 3.2
- openFABMap (Included as third party)
- Sophus (Included as third party)
When built there will be two executables in the ./bin directory
- sample_app : For use with the webcam
- main_on_images : For use with dataset images (download from TUM page mentioned above)
There is a camera calibration application in the camera_calibration directory. Just run the script build.sh and it will create an executable by name calibration in the same directory.