Skip to content

Beluga vision

Christopher Strøm edited this page Aug 22, 2021 · 1 revision

Intro

Beluga uses a Stereolabs ZED 2 stereo camera for all of its image acquisition. This camera comes with a powerful SDK that supports depth sensing, mapping etc. Stereolabs has also made a ROS wrapper that we use to integrate the camera in Belugas ROS stack.

Using the camera in ROS

The camera can be run through ROS by using some simple commands. The Xavier has all prerequisites installed.

Step by step:

  1. Source the zed-ros-wrapper workspace
  2. Start the ZED 2 node: roslaunch zed_wrapper zed2.launch
  3. Get all published topics with rostopic list. For example, the topics left/image_rect_color and right/image_rect_color include rectified color images from both sensors. These can be visualized in RViz e.g. List of all topics can be seen here.
  4. All published topics can be rosbagged easily: rosbag record <topic_name1> <topic_name2>

To edit the camera's parameters before launching the ZED 2 node, you may edit the .yaml files in ~/zed-ros-wrapper/zed_wrapper/params; namely common.yaml and zed2.yaml. This allows changing resolution, frame rate, depth sensing mode and more.

For more details, you may visit Stereolabs and ROS.

Custom camera applications

Stereolabs has created API's for Python and C++ that build upon the ZED SDK. These can be used to build custom camera applications that benefit from the features of the SDK.

Clone this wiki locally