AprilTagVision is a drop-in replacement for reacTIVision, using the Apriltags3 library developed by AprilRobotics. The AprilTag library and its fiducial markers, are faster, smaller, and more robust. AprilTagVision uses pupil-apriltags.

Download Mac binaries from https://github.com/dantaeyoung/AprilTagVision/tags.

For Mac: Extract apriltagvision_mac.zip. Run ./apriltagvision --help from the command line.

1. Download this repo.

2. Install the dependencies listed in the requirements.txt file: pip install -r requirements.txt

Basic:

• python apriltagvision.py

Select camera:

• python apriltagvision.py --camera 2

Select camera and force camera dimensions:

• python apriltagvision.py --camera 3 --width 640 --height 480

Select camera and output verbose debug output:

• python apriltagvision.py --camera 1 --verbose

Press p to pause processing. Press q to quit.

Like reacTIVision, AprilTagVision follows the TUIO Protocol 1.1 and sends messages out via OSC on port 3333.

Example messages might look like:

{"address":"/tuio/2Dobj","args":["source","apriltagvision"]}
{"address":"/tuio/2Dobj","args":["alive",100,101]}
{"address":"/tuio/2Dobj","args":["set",101,49,0.3847601115703583,0.32317110896110535,0.48030000925064087,0,0,0,0,0]}
{"address":"/tuio/2Dobj","args":["set",102,73,0.6303706765174866,0.22211359441280365,1.68340003490448,0,0,0,0,0]}
{"address":"/tuio/2Dobj","args":["fseq",78]}

Right now, AprilTagVision only supports fiducial markers. There is no camera calibration or touch detection.

tuioProcessor.py is a file that listens to TUIO messages on port 3333, processes them, and outputs 'tag appeared/moved/disappeared' signals to 1) via OSC on port 3334, and 2) via websockets on ws://localhost:8765.

Example messages might look like:

{"tagid": 27, "change": "appeared", "xpos": 0.5228530764579773, "ypos": 0.8543822169303894, "ang": 3.0566999912261963, "xvel": 0, "yvel": 0, "angvel": 0}
{"tagid": 27, "change": "moved", "xpos": 0.5225427150726318, "ypos": 0.8544207215309143, "ang": 3.0018999576568604, "xvel": 0, "yvel": 0, "angvel": 0.05480000004172325}
{"tagid": 27, "change": "disappeared"}

1. Run AprilTagVision with python apriltagvision.py --camera 0
2. While the above is running, in another terminal window, also run python tuioProcessor.py
3. In a browser window, open tuio_websocket_demo_3_table_simulate.html (or the other HTML files).

You should see

This code is currently only tested on a Mac M1. Please let me know if you try it on another OS or system!

Happy Hacking!