Its based on both the image_tools cam2image demos for ROS2 as well as the libuvc and usb_cam project for ROS1.
Features
- CameraInfo available
- CompressedImage topic (see compressed images for republishing using image_transport)
- Image topic
- Select camera (running the node for each camera connected)
There might be major changes to the code as it is a WIP. This is a simple camera driver using OpenCV. With this comes less flexibility for custom camera settings etc but simple to setup and use. If you have more complex requirements I've listed some alternate packages in the Alternate packages section.
/camera_info- topic for camera info/image_raw- topic for raw image data
Make sure to run setup.bash and local_setup.bash for all dependencies or symlink them into the repo.
Run
colcon build
ros2 run usb_camera_driver usb_camera_driver_node __ns:=/<your namespace> __params:=config.yaml
Available parameters:
frame_id-> transform frame_id of the camera, defaults to "camera"image_width-> image width (1280 default)image_height-> image height (720 default)fps-> video fps (10 default)camera_id-> id of camera to capture (0 - 100, 0 default)
To use the camera info functionality you need to load a file from the camera_calibration (https://github.com/ros-perception/image_pipeline/tree/ros2/camera_calibration) library and put it in/name it file:///Users/<youruser>/.ros/camera_info/camera.yaml
To get compressed images (works seamlessly with web streaming) republish the topic using image_transport which is available for ROS2.
ros2 run image_transport republish raw in:=image_raw compressed out:=image_raw_compressed
Make sure to link/install https://github.com/ros-perception/image_transport_plugins/tree/ros2 before to enable compressed image republishing using image_transport since its not included in the base package. More information here http://wiki.ros.org/image_transport, here http://wiki.ros.org/compressed_image_transport and here https://answers.ros.org/question/35183/compressed-image-to-image/
Other camera packages in ROS2 now available:
https://github.com/ros-perception/vision_opencv/tree/ros2