Example
We provided example code snippets for running OpenVSLAM with variety of datasets.
SLAM with Video Files
Tracking and Mapping
We provide an example snippet for using video files (e.g. .mp4) for visual SLAM.
The source code is placed at ./example/run_video_slam.cc.
The following options are allowed:
$ ./run_video_slam -h
Allowed options:
-h, --help produce help message
-v, --vocab arg vocabulary file path
-m, --video arg video file path
-c, --config arg config file path
--mask arg mask image path
--frame-skip arg (=1) interval of frame skip
--no-sleep not wait for next frame in real time
--auto-term automatically terminate the viewer
--debug debug mode
--eval-log store trajectory and tracking times for evaluation
-p, --map-db arg store a map database at this path after SLAM
.yaml) according to the camera parameters.orb_vocab.dbow2 in the zip file.Localization
We provide an example snippet for using video files (e.g. .mp4) for localization based on a prebuilt map.
The source code is placed at ./example/run_video_localization.cc.
The following options are allowed:
$ ./run_video_localization -h
Allowed options:
-h, --help produce help message
-v, --vocab arg vocabulary file path
-m, --video arg video file path
-c, --config arg config file path
-p, --map-db arg path to a prebuilt map database
--mapping perform mapping as well as localization
--mask arg mask image path
--frame-skip arg (=1) interval of frame skip
--no-sleep not wait for next frame in real time
--auto-term automatically terminate the viewer
--debug debug mode
.yaml) according to the camera parameters.orb_vocab.dbow2 in the zip file.You can create a map database file by running one of the run_****_slam executables with --map-db map_file_name.msg option.
SLAM with Image Sequences
Tracking and Mapping
We provided an example snippet for using image sequences for visual SLAM.
The source code is placed at ./example/run_image_slam.cc.
The following options are allowed:
$ ./run_image_slam -h
Allowed options:
-h, --help produce help message
-v, --vocab arg vocabulary file path
-i, --img-dir arg directory path which contains images
-c, --config arg config file path
--mask arg mask image path
--frame-skip arg (=1) interval of frame skip
--no-sleep not wait for next frame in real time
--auto-term automatically terminate the viewer
--debug debug mode
--eval-log store trajectory and tracking times for evaluation
-p, --map-db arg store a map database at this path after SLAM
.yaml) according to the camera parameters.orb_vocab.dbow2 in the zip file.Localization
We provided an example snippet for using image sequences for localization based on a prebuilt map.
The source code is placed at ./example/run_image_localization.cc.
The following options are allowed:
$ ./run_image_localization -h
Allowed options:
-h, --help produce help message
-v, --vocab arg vocabulary file path
-i, --img-dir arg directory path which contains images
-c, --config arg config file path
-p, --map-db arg path to a prebuilt map database
--mapping perform mapping as well as localization
--mask arg mask image path
--frame-skip arg (=1) interval of frame skip
--no-sleep not wait for next frame in real time
--auto-term automatically terminate the viewer
--debug debug mode
.yaml) according to the camera parameters.orb_vocab.dbow2 in the zip file.You can create a map database file by running one of the run_****_slam executables with --map-db map_file_name.msg option.
SLAM with Standard Datasets
KITTI Odometry dataset
KITTI Odometry dataset is a benchmarking dataset for monocular and stereo visual odometry and lidar odometry that is captured from car-mounted devices.
We provided an example source code for running monocular and stereo visual SLAM with this dataset.
The source code is placed at ./example/run_kitti_slam.cc.
Start by downloading the dataset from here.
Download the grayscale set (data_odometry_gray.zip).
After downloading and uncompressing it, you will find several sequences under the sequences/ directory.
$ ls sequences/
00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21
In addition, download a vocabulary file for DBoW2 from here and uncompress it.
You can find orb_vocab.dbow2 in the zip file.
A configuration file for each sequence is contained under ./example/kitti/.
If you built examples with Pangolin Viewer support, a map viewer and frame viewer will be launced right after executing the following command.
# at the build directory of OpenVSLAM
$ ls
...
run_kitti_slam
...
# monocular SLAM with sequence 00
$ ./run_kitti_slam \
-v /path/to/orb_vocab/orb_vocab.dbow2 \
-d /path/to/KITTI/Odometry/sequences/00/ \
-c ../example/kitti/KITTI_mono_00-02.yaml
# stereo SLAM with sequence 05
$ ./run_kitti_slam \
-v /path/to/orb_vocab/orb_vocab.dbow2 \
-d /path/to/KITTI/Odometry/sequences/05/ \
-c ../example/kitti/KITTI_stereo_04-12.yaml
The following options are allowed:
$ ./run_kitti_slam -h
Allowed options:
-h, --help produce help message
-v, --vocab arg vocabulary file path
-d, --data-dir arg directory path which contains dataset
-c, --config arg config file path
--frame-skip arg (=1) interval of frame skip
--no-sleep not wait for next frame in real time
--auto-term automatically terminate the viewer
--debug debug mode
--eval-log store trajectory and tracking times for evaluation
-p, --map-db arg store a map database at this path after SLAM
EuRoC MAV dataset
EuRoC MAV dataset is a benchmarking dataset for monocular and stereo visual odometry that is captured from drone-mounted devices.
We provide an example source code for running monocular and stereo visual SLAM with this dataset.
The source code is placed at ./example/run_euroc_slam.cc.
Start by downloading the dataset from here.
Download the .zip file of a dataset you plan on using.
After downloading and uncompressing it, you will find several directories under the mav0/ directory.
$ ls mav0/
body.yaml cam0 cam1 imu0 leica0 state_groundtruth_estimate0
In addition, download a vocabulary file for DBoW2 from here and uncompress it.
You can find orb_vocab.dbow2 in the zip file.
We provided the two config files for EuRoC, ./example/euroc/EuRoC_mono.yaml for monocular and ./example/euroc/EuRoC_stereo.yaml for stereo.
If you have built examples with Pangolin Viewer support, a map viewer and frame viewer will be launched right after executing the following command.
# at the build directory of OpenVSLAM
$ ls
...
run_euroc_slam
...
# monocular SLAM with any EuRoC sequence
$ ./run_euroc_slam \
-v /path/to/orb_vocab/orb_vocab.dbow2 \
-d /path/to/EuRoC/MAV/mav0/ \
-c ../example/euroc/EuRoC_mono.yaml
# stereo SLAM with any EuRoC sequence
$ ./run_euroc_slam \
-v /path/to/orb_vocab/orb_vocab.dbow2 \
-d /path/to/EuRoC/MAV/mav0/ \
-c ../example/euroc/EuRoC_stereo.yaml
The following options are allowed:
$ ./run_euroc_slam -h
Allowed options:
-h, --help produce help message
-v, --vocab arg vocabulary file path
-d, --data-dir arg directory path which contains dataset
-c, --config arg config file path
--frame-skip arg (=1) interval of frame skip
--no-sleep not wait for next frame in real time
--auto-term automatically terminate the viewer
--debug debug mode
--eval-log store trajectory and tracking times for evaluation
-p, --map-db arg store a map database at this path after SLAM
TUM RGBD dataset
Will be written soon.
The following options are allowed:
$ ./run_tum_slam -h
Allowed options:
-h, --help produce help message
-v, --vocab arg vocabulary file path
-d, --data-dir arg directory path which contains dataset
-a, --assoc arg association file path
-c, --config arg config file path
--frame-skip arg (=1) interval of frame skip
--no-sleep not wait for next frame in real time
--auto-term automatically terminate the viewer
--debug debug mode
--eval-log store trajectory and tracking times for evaluation
-p, --map-db arg store a map database at this path after SLAM
SLAM with UVC camera
Tracking and Mapping
We provided an example snippet for using a UVC camera, which is often called a webcam, for visual SLAM.
The source code is placed at ./example/run_camera_slam.cc.
The following options are allowed:
$ ./run_camera_slam -h
Allowed options:
-h, --help produce help message
-v, --vocab arg vocabulary file path
-n, --number arg camera number
-c, --config arg config file path
--mask arg mask image path
-s, --scale arg (=1) scaling ratio of images
-p, --map-db arg store a map database at this path after SLAM
--debug debug mode
-n option..yaml) according to the camera parameters.-s option. Please modify the config accordingly.orb_vocab.dbow2 in the zip file.Localization
We provided an example snippet for using a UVC camera for localization based on a prebuilt map.
The source code is placed at ./example/run_camera_localization.cc.
The following options are allowed:
$ ./run_camera_localization -h
Allowed options:
-h, --help produce help message
-v, --vocab arg vocabulary file path
-n, --number arg camera number
-c, --config arg config file path
--mask arg mask image path
-s, --scale arg (=1) scaling ratio of images
-p, --map-db arg path to a prebuilt map database
--mapping perform mapping as well as localization
--debug debug mode
-n option..yaml) according to the camera parameters.-s option. Please modify the config accordingly.orb_vocab.dbow2 in the zip file.