Using AprilTag python library on Windows 10?

Question

I'm in need of a fiducial marker for a rover project as its primary control is image processing for steering down a driveway and when it gets to the end i need it to go to a precise location.

I was going to use ArUco tags but my version of OpenCV for the OpenVino framework doesn't support it. Then i tried to use AprilTags but i couldn't get it to install on my Windows 10 machine and later discovered its not supported on that OS. While my robot uses RaspPi i absolutely need my desktop environment for development and debugging.

Anyone get AprilTag library compiled for Windows Python 3? If not does anyone know of an alternate tag library (not ArUco) that is similar? I don't need anything fancy here - just something printed on a 4" x 4" (ish) piece of paper that my image processing pipeline can easily pick out of the image and i can then navigate to it. I don't even need to encode data on it - just need something unique to lock into.

Much appreciated!

Answer 1

Intel® Distribution of OpenVINO™ Toolkit can also be installed through PyPI repository instead of installing the binary package that had compiled OpenCV community version.

• Runtime package with the Inference Engine inside.
• Developer package that includes the runtime package as a dependency, Model Optimizer, Accuracy Checker and Post-Training Optimization Tool.

Next, you can get the ArUco library that included in the opencv-contrib-python package which you can install via:

pip install opencv-contrib-python

Answer 2

For anyone who still needs an AprilTag Python 3 library for Windows 10/11, my team used the pyAprilTag library create by Chen Feng and Ruoyu Wang. Here is the link to the Github Repo:

https://github.com/ai4ce/pyAprilTag

Even though you will see it on PyPI, you are not able to install it with the "pip install pyAprilTag" or "pip3 install pyAprilTag" commands anymore. You will need to follow the instructions from the repo itself to install it. Additionally, you will see in the repo that the easiest way to install it is with a combination of pip (or pip3) and conda. I am sorry if that is a deal breaker for anyone.

Extra Notes: The repo says that you can install it on MacOS by compiling from source, but our team has not been able to do that yet. That does not mean that it is impossible, per say, we just gave up quickly since we already had another Python library for MacOS:

https://github.com/swatbotics/apriltag

We ended up just doing conditional imports for compatibility with all the operating systems:

https://github.com/cornell-cup/cs-reminibot/blob/vision_doc_updated_gui_interpolation/vision/apriltag-py/python/detector.py

Lastly, our team ended up running the step 1 ("try to create and activate a new conda environment with") commands in the same directory that we downloaded the "opencv_py37_conda-forge.txt" file. The original steps were not working for us, but it could have just been an issue with the computer we were using at the time.

Anyways, best of luck with your endeavors!