Simple RGB-D SLAM Implementation for Research (under 1500 lines)
Dependencies are listed in the table below along with the version used during development and testing.
Dependency | Tested Version |
---|---|
OpenCV | 4.5.5 |
Ceres | 2.0.0 |
Sophus | 1.0.0 |
Eigen | 3.3.4 |
CSparse | 5.1.2 |
Pangolin | 0.5 |
Glog | 0.3.5 |
DBoW3 | 1.0 |
git clone https://github.com/93won/RGBD_SLAM && cd RGBD_SLAM
mkdir build && cd build
make
[1] Download a sequence from https://vision.in.tum.de/rgbd/dataset/freiburg2/rgbd_dataset_freiburg2_desk.tgz and uncompress it.
[2] Execute associate.py in the script folder as bellow to create the associations.txt file.
python associate.py DATA_FOLDER/rgb.txt DATA_FOLTER/depth.txt
[3] Download DBoW3 ORB Vocabulary file from https://github.com/rmsalinas/DBow3/blob/master/orbvoc.dbow3
[4] Edit data directory in config/f2_desk.yaml
[5] Run ./bin/test_rgbd