Add ArUco families support

[fspindle]

This PR introduces native support for ArUco tag families.

Integrated the ArUco dictionaries into the AprilTag ecosystem. This allows users to detect ArUco markers using the robust AprilTag 3 detector without needing external conversion tools.

• Added families: 4x4, 5x5, 6x6, 7x7 (various sizes from 50 to 1000) and ArUcoMIP36h12.
• Updated apriltag_pywrap, apriltag_demo, and opencv_demo to support these new families.
• Added comprehensive tests in test_quick_decode.c.

This PR replaces #428 that was closed.

[christian-rauch]

How are the "tagAruco" files generated? Since this adds quite a few files to the repo, maybe we should add a new "tag" subfolder and place them there?

[fspindle]

This work was originally carried out by @s-trinh
Information for generating ArUco codes can be found here:

• Advice: how to decode ArUco codeword with AprilTag code? #374 (comment)
• Add support for ArUco 7x7 tag family lagadic/visp#1835

The code provided in this issue for generating the 36h11 family was used.

I agree that moving all tag*.{c,h} files in a dedicated tag/ folder could be a good idea.

[fspindle]

maybe we should add a new "tag" subfolder and place them there?

Adding a tag folder breaks backward compatibility with previous releases. Users will need to update their code from #include <tag36h11.h> to #include <tag/tag36h11.h>. Is this breaking change intentional?

[christian-rauch]

maybe we should add a new "tag" subfolder and place them there?

Adding a tag folder breaks backward compatibility with previous releases. Users will need to update their code from #include <tag36h11.h> to #include <tag/tag36h11.h>. Is this breaking change intentional?

Yes, I wouldn't do this for the old tags as this breaks the paths as you suggested. We can leave these 18 files, or at least the headers where they are currently. But because you add many more files, I would prefer to see them in a subfolder. Maybe let's call this just aruco to distinguish it from the other tags.

[fspindle]

@christian-rauch Done

[christian-rauch]

Can you address the reported compilation errors in the CI and squash your fixes into the commits again?

One of the reported errors is:

/__w/apriltag/apriltag/test/test_quick_decode.c:15:10: fatal error: tagAruco4x4_50.h: No such file or directory
   15 | #include "tagAruco4x4_50.h"
      |          ^~~~~~~~~~~~~~~~~~

[christian-rauch]

Can you squash your fixes into the commit that caused the issue in the first place?