ros2-groundgrid

ROS 2 support of GroundGrid: dcmlr/groundgrid.

Paper: GroundGrid: LiDAR Point Cloud Segmentation and Terrain Estimation; IEEE'2024

Requirements

• ROS 2 Humble

Build & Run

Build

rosdep update
rosdep install -y --from-paths . --ignore-src --rosdistro $ROS_DISTRO
colcon build --symlink-install --cmake-args -DCMAKE_BUILD_TYPE=Release
source install/setup.bash

Run

ros2 launch groundgrid groundgrid.launch.xml

Sample Dataset

Download Semantic KITTI Dataset.

Downloaded dataset structure is as follows:

ros-groundgrid
├── data
│   └── kitti
│       ├── dataset
│       |   └── sequences
|       │       ├── 00
│       │       │   ├── calib.txt
│       │       │   ├── labels
│       │       │   ├── poses.txt
│       │       │   ├── times.txt
│       │       │   └── velodyne
│       │       ├── 01
...

Requirements

ros-humble-rosbag2-storage-mcap is required to serialize bag files into MCAP format.

sudo apt install ros-humble-rosbag2-storage-mcap

Save ROS Bag

Run kitti2bag.py to convert dataset to bag file in mcap format:

python3 scripts/kitti2bag.py </KITTI/DATASET/PATH>

For example in the above case:

python3 scripts/kitti2bag.py ./data/kitti/dataset

In order to convert the particular sequence, please specify -s <SEQUENCE_NUMBER>:

python3 scripts/kitti2bag.py </KITTI/DATASET/PATH> -s 00

Play Bag file

After finishing conversion, you can run the recorded bag file as below:

ros2 bag play -s mcap </KITTIT/BAG/DIR>

Bag files will be saved into data/kitti_bag/semantic_kitti_<SEQUENCE> by default:

ros2 bag play -s mcap ./data/kitti_bag/semantic_kitti_00

Visualize

For visualization, rviz/groundgrid.rviz is prepared for the rviz configuration by default:

rviz2 -d ./rviz/groundgrid.rviz