RGBD SLAM - Building A ROS 2 Docker Container For Object Detection

Dockerfile

• I installed jupyter notebook so I can experiment with it remotely.

1
2
3
4
# -U is to upgrade to the latest version
RUN python3 -m pip install -U jupyter notebook 
# jupyter notebook port
EXPOSE 8888

Docker Runtime Args

• --runtime=nvidia: enable Nvidia Container Runtime, a “runtime” that connects Nvidia GPU with docker. If your laptop doesn’t have an Nvidia GPU, simply remove this arg.
• -p 8888 to expose the Jupyter Notebook Port
• jupyter notebook --ip=0.0.0.0 --no-browser --allow-rootfollows the image name. This is the command to run. without -d, docker will run this process in the foreground, which will terminates the container upon SIGINT (ctrl-C)
• Make sure /dev/shm is shared properly
  • On the same machine, ROS2 uses the directory for messaging
  • But this wouldn’t work -d /dev/shm:/dev/shm
  • use -v instead

Push To Docker Hub

1
2
3
docker login  #put in docker creds
docker tag <MY_IMAGE>:latest <MY_USERNAME>/<MY_IMAGE>:latest
docker push <MY_USERNAME>/<MY_IMAGE>:latest

• Docker Hub requires images to be pushed to the repo that corresponds to our Docker Hub username. So, we need to tag the image with the proper username.

Machine Learning Docker on Jetson

To fully unleash the power of an Nvidia Jetson Machine for machine learning, there are multiple ways to set up the environment. A common way is to set up conda. In my application though, I prefer to use an Docker image so I can hook up my ML application with ROS / ROS2 applications. To do that, we need to set up a CUDA image properly, then select the correct pytorch version.

Luckily, dusty-nv and others have created a nice Github package to streamline this process. This package is a small Docker build system that creates the final desired Docker image through multi-stage builds (like a chain). E.g., one can build an image with pytorch, ROS Noetic and Jupyterlab.

If you don’t want much hassle like me, this could be a good starting point. But just in case you are curious, here is the place to check for the Pytorch version

For flashing the Jetson itself, see Flashing the Orin Nano.

• Previous
  RGBD SLAM - What Deteriorates Its Performance
• Next
  Robotics - [Jetson 1] Flashing the Orin Nano (SD Card and NVMe)