peak_cam

A Linux ROS C++ Node that wraps the driver API for IDS vision cameras using IDS peak software. Tested on Ubuntu 18.04 LTS and 20.04 LTS.

Supports the set_camera_info service used by the ROS camera_calibration package.

How to run

Before running the code

1. Install ROS
2. Install IDS peak
   1. Install dependencies: sudo apt install libqt5core5a libqt5gui5 libqt5widgets5 libqt5quick5 qml-module-qtquick-window2 qml-module-qtquick2 qml-module-qtquick-dialogs qml-module-qtquick-controls qml-module-qtquick-layouts qml-module-qt-labs-settings qml-module-qt-labs-folderlistmodel libusb-1.0-0
   2. Install package: sudo dpkg -i ids-peak-<version>-<arch>.deb
   3. (Fix OpenCV version issue if working on NVIDIA Jetson: sudo ln -s /usr/include/opencv4/opencv2/ /usr/include/opencv)
3. Test connection to your camera using ids_visioncockpit
4. (Update camera firmware if needed: ids_deviceupdate -s *<last-four-digits-serialnumber> -U --guf <path-to-guf-file>)

Configuration for GigE cameras

1. Enable jumboframes for the ethernet interface(s) (e.g., eth0) used for your camera(s) by setting MTU to 9000 in your network manager, by editing /etc/network/interfaces or non-permanent using ip link set dev eth0 mtu 9000
2. Increase receive buffer size as recommended in the IDS manual: sudo /usr/local/scripts/ids_set_receive_buffer_size.sh
3. Make sure to adjust DeviceLinkThroughputLimit in the .yaml configuration file according to your desired framerate and available hardware (employing separate network interfaces may be beneficial)

Running the code

1. Clone the repository to your Linux computer

2. Generate a ROS workspace

   $ mkdir -p camera_ws/src/

3. Copy the peak_cam package into your ROS workspace and build it

   $ cp -r peak_cam/ camera_ws/src/

   $ cd camera_ws/ && catkin_make && source devel/setup.bash

4. Set parameters such as ROS topic and acquisition rate under launch/params/peak_cam_params.yaml

5. Plug the IDS vision camera and launch the node

   $ roslaunch peak_cam peak_cam_node.launch

6. Stop the node with Ctrl-C (SIGINT) for controlled shutdown

For multiple cameras, create a .launch and a .yaml file for each camera.

Hint: Sometimes the cameras are only accesible as root. Try sudo -s in your terminal and launch the node again.

Triggering for stereo vision

To synchronously acquire images from two cameras in a stereo vision setup, you might want to let one camera trigger the other.

1. Setup the GPIO wiring as proposed by IDS according to the pin layout found in your camera's datasheet (example color coding for GV-5270FA):

   

2. Configure the primary (master) camera's .yaml file with

   Line1Source: "ExposureActive"
   TriggerSource: "Off"
   

   and the secondary camera's .yaml with

   Line1Source: "Off"
   TriggerSource: "Line0"
   

Trigger using external pulses

The cameras can also be triggered by the pulses of external devices such as a lidar sensor.

1. (Activate the trigger output of your external device, e.g., using the multipurpose_io_mode flag for Ouster lidars as discussed in section 5.2.2 of their software user manual)

2. Setup the GPIO wiring in compliance to your hardware, e.g., with an optoisolated open collector sync pulse circuit:

   

3. Configure your camera's .yaml files with

   Line1Source: "Off"
   TriggerSource: "Line0"
   

Copyright (c) 2020, Sherif Nekkah, Felix Keppler (Fraunhofer IVI), Johannes Schäfer (Fraunhofer IVI) and Contributors

All rights reserved.

BSD license: see LICENSE file