Skip to content

pointcloud2gazebo is a package that allows you to create a Gazebo World from a 3D point cloud.

License

Notifications You must be signed in to change notification settings

CyberAgentAILab/pointcloud2gazebo

Repository files navigation

pointcloud2gazebo

Introduction

pointcloud2gazebo is a package that allows you to create a Gazebo World from a 3D point cloud.

Input Point Cloud Generated Mesh Gazebo World
input point cloud generated mesh Gazebo World

Demo Movie

The following demo showcases a robot navigating a Gazebo World generated by this package.

Demo Movie

Requirements

  • python ==3.10
  • Open3D
  • NumPy
  • Traitlets

Preparation

Installing the Package

pip install -r requirements.txt

If you use Docker for setup of this tool, please execute the following command.

docker build -t pointcloud2gazebo .
./launch_container.sh

Generating 3D Point Cloud Map

You will need to generate a 3D point cloud map. As an example, we have prepared data/cloud.ply. This data was generated using RTAB-Map.

Usage

Generating a Gazebo World

python3 generate_world.py --input=data/cloud.ply

If you enable to visualize mesh, please add --vis option.

python3 generate_world.py --input=data/cloud.ply --vis

If you encounter the following error, please execute xhost +local: before execute generate_world.py.

$ python3 generate_world.py --input=data/cloud.ply --vis
No protocol specified
[Open3D WARNING] GLFW Error: X11: Failed to open display :0
[Open3D WARNING] Failed to initialize GLFW
[Open3D WARNING] [DrawGeometries] Failed creating OpenGL window.

Copying the Generated Gazebo World to GAZEBO_MODEL_PATH

mkdir -p ~/.gazebo/models
cp -r models/pointcloud2gazebo ~/.gazebo/models/

Launching the Gazebo World

If you try to spawn Gazebo World generated by this package, you need to install Gazebo. Please read official document.

gazebo worlds/pointcloud2gazebo.world

Appendix

Configuration

You can customize the package using config/config.py.

Parameter Description Type Unit
Crop.min_z Minimum height of points to create Gazebo world. float meter
Crop.max_z Maximum height of points to create Gazebo world. float meter
NoiseRemoval.nb_points Number of points within the radius. Refer to open3d.geometry.PointCloud.remove_radius_outlier. int -
NoiseRemoval.radius Radius of the sphere. Refer to open3d.geometry.PointCloud.remove_radius_outlier. float meter
VoxelDownSample.voxel_size Voxel size to downsample into. Refer to open3d.geometry.PointCloud.voxel_down_sample. float meter
TriangleMesh.alpha Parameter to control the shape. A large value will yield a shape close to the convex hull. Refer to open3d.geometry.TriangleMesh.create_from_point_cloud_alpha_shape. float -

The sample data (data/cloud.ply) was generated using RTAB-Map. The following parameters were set for this data:

c.Crop.min_z = -1.2
c.Crop.max_z = 1.0

About

pointcloud2gazebo is a package that allows you to create a Gazebo World from a 3D point cloud.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published