fiducialary 1.1.0

Generate unique circular fiducial markers for use in vision systems such as photogrammetry.

fiducialary

fiducialary is a tool for generating fiducial marks for use in imaging systems. These artificial markers made using circular codes can be used in photogrammetry, volumetric capture and vision search.

This project also provides some pre-generated markers you can use without having to run the program to generate them.

Contributors

• Luke Miller

Thanks

fiducialary made for 'Project 39', a creative art project supported by Creative Victoria.

Using existing markers

Looking for markers to use?

The marker images generated by this program in the markers folder are free to use in any way, commercial or otherwise. I release these free fiducial markers into the public domain. Download them direct: https://gitlab.com/dodgyville/fiducialary/tree/master/markers

Note: The fiducialary program that generated them is licensed under the open source LGPLv3 and is not public domain.

Recommended: Print out one of the A4 pdf files and separate the markers for use.

How to use markers

• Markers work best when surrounded by space - so don't bend, cut or write on markers.
• Try to have three or four markers in each image.
• Do not use duplicate markers, all markers should be unique.
• Some research indicates that the more of your pixels occupied by a marker the better but real-world results suggest not to make them too large.

Generating your own markers

Install

pip install fiducialary

Source

git clone https://gitlab.com/dodgyville/fiducialary

Usage

Generate a PDF of markers

From the command line, run:

fiducialary

it will output a PDF

A single marker from inside python

from fiducialary import generate_marker 

marker = generate_marker()
marker.write_to_png(f"example.png")

A batch of markers

from fiducialary import generate_markers 

markers = generate_markers()
save_markers_individually(markers)

A 20-bit marker

marker = generate_marker(bits=20)

Example marker

Known issues

• When generating multiple markers, duplicate patterns may occur but rotated compared to each other. So watch out for that.

Roadmap

• Test for rotated duplicates
• Better documentation

Changelog

v1.1.0

• switch to pyproject and poetry
• install as command line tool available at fiducialary

v1.0.1

• fix packaging issue

v1.0.0

• initial release