SHOP News & Articles Contact Impressum

[Unitree Go1] Guide for the UnitreeCameraSDK


My deep dive into Unitrees repositories brought me to the last one: The UnitreeCameraSDK.

The question is, what is it capable of and how do I use it correctly?

I’ve tried to run the given examples, but they simply do not work out of the box.
The errors I get are listed in my opened issue, which is yet to be answered. But I don’t have any hopes Unitree is going to answer, since this repository doesn’t seem to be up on their priority list.

Is it even worth it for me to spend time getting it to run?

I would love to hear from someone, who has experiences with the UnitreeCameraSDK!

Thank you in advance


I’ve found some advice to kill some processes before starting the examples:

In advance, kill the camera & imageai process.

ps aux | grep camera
sudo ~/Unitree/camerarosnode/cameraRosnode/ 
ps aux | grep camera
ps aux | grep image
sudo kill <PSID>
ps aux | grep image

But this means I’d have to kill them every time I rebooted the robot, which would be quite unpleasant.

Hi @jgrube
A detailed discussion can be found here for killing the processes and how to start.
Actually yes you have to kill the scripts occupying cameras. Because ROS is not the primary software which is meant to run on the robot. You can check in which startup scripts cameras are occupied and disable them or you can run the kill script which is inside camera ros node package at startup which will kill everything immediately.

For your other question of using sdk for ros package we have a very small experimental package at the moment which you can download from the same link which I shared you previously but make sure the node to run on respective NVIDIA.

To build camera driver package you might have to install some packages which you can install using apt-offline

In case any issue please feel free to post.

Thanks, for the background info!
The issue you linked to is actually my own (which I mentioned above) ^^

Footer With Address And Phones