introlab3it/rtabmap_ros repository overview

⁠rtabmap_ros

ROS wrapper for RTAB-Map⁠ — graph-based SLAM with appearance-based loop closure detection. Builds and maintains a 3D map from RGB-D, stereo or lidar data, and exports it as an occupancy grid, a point cloud or an OctoMap.

Source, documentation and issues → github.com/introlab/rtabmap_ros⁠

⁠Tags

ROS 2 — built from the ros2⁠ branch

ROS 1 — built from the master⁠ branch. Only the Noetic images are still rebuilt on ROS 1 changes; Melodic and Kinetic remain available but are no longer updated.

<distro> is what apt install ros-<distro>-rtabmap-ros gives you — the versions released on the ROS build farm.

<distro>-latest is built from the current rtabmap and rtabmap_ros sources, including optional dependencies that have no ROS binaries, and with multi-camera synchronization enabled.

⁠Quick start

Run the SLAM node against an RGB-D camera and an odometry source already publishing on the host:

docker run -it --rm \
  --user $UID -e ROS_HOME=/tmp/.ros -e OMP_WAIT_POLICY=passive \
  --network=host --ipc=host \
  -v ~/.ros:/tmp/.ros \
  introlab3it/rtabmap_ros:jazzy-latest \
  ros2 run rtabmap_slam rtabmap -d --ros-args \
    -p subscribe_depth:=true \
    -p frame_id:=camera_link \
    -p database_path:=/tmp/.ros/rtabmap.db \
    -r rgb/image:=/camera/color/image_raw \
    -r depth/image:=/camera/depth/image_rect_raw \
    -r rgb/camera_info:=/camera/color/camera_info \
    -r odom:=/odom

-d deletes the previous database on start. The map is written to ~/.ros/rtabmap.db on the host.

This node does the mapping only — it expects odometry on odom. To bring up odometry and SLAM together, use ros2 launch rtabmap_launch rtabmap.launch.py instead.

Running the rtabmap_viz GUI needs X11 and GPU access — see docker/README.md⁠ for that and for more examples.

⁠License

BSD-3-Clause. RTAB-Map itself may be built with components under other licenses; see the rtabmap⁠ repository.