Point Transformation

ROS2 Node for 2D / 3D Point Transformation

How to start:

1. Set the ROS_DOMAIN_ID in ./.env

   ROS_DOMAIN_ID=66

2. Adapt parameters for specific use case in ros/config/params.yaml
3. (only once) Build the docker container

   . build_docker.sh

4. Start the docker container

   . start_docker.sh

Interface:

Point Transformation Node

• point_transformation_node (ROS2 Dashing in C++)

Services

• point_transformation_node/pixel_to_point (Custom type)

  geometry_msgs/Point[] pixels        # pixels in rgb image
  uint32 height 0                     # rgb image height, default if unassigned = 0
  uint32 width 0                      # rgb image width, default if unassigned = 0
  sensor_msgs/Image depth_image       # corresponding depth image
  string camera_type                  # used camera for depth image ("roboception" or "realsense")
  ---
  geometry_msgs/Point[] points        # xyz of points in [m]

Parameter

Change in ros/config/params.yaml

opening_angle_horizontal: 61.0  # Horizontal opening angle of the used Camera (Roboception) in degrees
opening_angle_vertical: 48.0    # Vertical opening angle in degrees
focal_factor: 0.75              # Focal factor of the camera lense
width: 1920                     # Default width if no image width is given to the service
height: 1080                    # Default height
default_depth: 1.0              # Default depth in m if no depth image is given
max_pixel_range_for_depth_matching: 5 # Search in a NxN area around the correct pixel if it is nan

How to launch:

ros2 launch point_transformation point_transformation.launch.py

Testing:

ros2 run point_transformation client.py

Todo:

• update .svg with correct interface