robotraconteur-camera-driver

Robot Raconteur Camera Driver

The robotraconteur-camera-driver package provides a camera driver for Robot Raconteur based on OpenCV video capture. See the OpenCV documentation for more information on the video capture system and its capabilities. The driver allows for selecting the camera by index, and changing some of the parameters made available by OpenCV. OpenCV does not provide a mechanism to enumerate cameras and their capabilities so there is no option to list available cameras.

This driver uses the com.robotraconteur.imaging.Camera standard type.

Install using Python pip (Windows, Mac, Linux) or using Docker (Linux).

Connection Info

The default connection information is as follows. These details may be changed using --robotraconteur-* command line options when starting the service. Also see the Robot Raconteur Service Browser to detect services on the network.

• URL: rr+tcp://localhost:59823?service=camera
• Device Name: camera or device name in the configuration file
• Node Name: com.robotraconteur.imaging.camera
• Service Name: camera
• Root Object Type:
  • com.robotraconteur.imaging.Camera

Installation

Install using Python pip

python -m pip install robotraconteur-camera-driver

On Linux use python3 instead of python. On Ubuntu the python3-pip apt package must be installed to use python pip.

See the Docker section for installation instructions using docker.

Usage

Example usage:

python robotraconteur_camera_driver.py --camera-info-file=config/generic_webcam_1080p_default_camera_info.yml --width=1280 --height=720 --fps=20 --device-id=0

On Ubuntu, use python3 instead of python

Command Line Options

The following command line arguments are available:

• --camera-info-file= - The camera info file. Info files are available in the config/ directory. See camera info file documentation. Also see camera_client_calibrate_intrinsic.py example.
• --device-id= - The Device ID used to identify the camera. Default is 0. Passed to the OpenCV Video Capture creation function.
• --width= - Captured image width (OpenCV CAP_PROP_FRAME_WIDTH)
• --height= - Captured image height (OpenCV CAP_PROP_FRAME_HEIGHT)
• --fps= - Frames per second (OpenCV CAP_PROP_FPS)
• --focus= - Camera focus (OpenCV CAP_PROP_FOCUS)
• --exposure= - Camera exposure (OpenCV CAP_PROP_EXPOSURE)
• --gain= - Camera gain (OpenCV CAP_PROP_GAIN)
• --brightness= - Camera brightness (OpenCV CAP_PROP_BRIGHTNESS)
• --contrast= - Camera contrast (OpenCV CAP_PROP_CONTRAST)
• --saturation= - Camera saturation (OpenCV CAP_PROP_SATURATION)

See the OpenCV Video Capture and OpenCV Video Capture Properties documentation for more information on these options.

The focus, exposure, gain, brightness, contrast, and saturation properties can be changed by the client using the setf_param function during runtime. All parameters are scalar double types. For example:

# camera_client connected using RRN.ConnectService
focus_value=10
camera_client.setf_param("focus", RR.VarValue(focus_value, "double"))

All Robot Raconteur node setup command line options are supported. See Robot Raconteur Node Command Line Options

By default the node listens on TCP port 59823 and have the NodeName com.robotraconteur.imaging.Camera. If there are multiple cameras, it will be necessary to change these two values for each additional camera. These can be overridden using Robot Raconteur node setup command line options. An additional camera info file with a different device name is also needed. For example,

python robotraconteur_camera_driver.py --camera-info-file=config/generic_webcam_1080p_default_camera_info2.yml --width=1280 --height=720 --fps=20 --device-id=1 --robotraconteur-tcp-port=54444 --robotraconteur-node-name=camera2

Examples

The following examples can be found in the examples/ directory:

• camera_client_capture_frame.py - Basic example to capture and display a single frame
• camera_client_capture_frame_compressed.py - Capture and display a frame from the camera using a compressed format. The compressed image is more efficient to transfer but requires additional computational power to compress the image. Lossless PNG is used to compress the image.
• camera_client_image.py - Stream full resolution uncompressed images
• camera_client_image_preview.py - Stream "preview" images, which are mjpeg compressed and reduced resolution for live preview use only
• camera_client_aruco_detection.py - Example of streaming detection of ArUco markers
• camera_client_capture_calibration_images.py and camera_client_calibrate_intrinsic.py - Examples capturing images and generating intrinsic calibration parameters. Outputs Yaml format compatible with info files.

Docker Usage

sudo docker run --rm --net=host --privileged -v /var/run/robotraconteur:/var/run/robotraconteur -v /var/lib/robotraconteur:/var/lib/robotraconteur wasontech/robotraconteur-camera-driver --camera-info-file=config/generic_webcam_1080p_default_camera_info.yml --width=1280 --height=720 --fps=20 --device-id=0

It may be necessary to mount a docker "volume" to access configuration yml files that are not included in the docker image. See the docker documentation for instructions on mounting a local directory as a volume so it can be accessed inside the docker.

License

License: Apache 2.0

Author: John Wason, PhD

Acknowledgment

This work was supported in part by the Advanced Robotics for Manufacturing ("ARM") Institute under Agreement Number W911NF-17-3-0004 sponsored by the Office of the Secretary of Defense. The views and conclusions contained in this document are those of the authors and should not be interpreted as representing the official policies, either expressed or implied, of either ARM or the Office of the Secretary of Defense of the U.S. Government. The U.S. Government is authorized to reproduce and distribute reprints for Government purposes, notwithstanding any copyright notation herein.

This work was supported in part by the New York State Empire State Development Division of Science, Technology and Innovation (NYSTAR) under contract C160142.