I hope that this can be useful to some teams out there who are struggling with writing their own vision processing code, everything is licensed under GPL. You can find all of this on our team’s github page. There is also a guide on how each script works, as well as tutorials on the basics of vision processing to help out complete beginners. Questions and criticism are more than welcome!
Repo for main vision code that’s on the robot: https://github.com/team401/2017-Vision-Code.
We’re using OpenCV in C++ on an NVIDIA Jetson TX1, comms to roboRIO via zeroMQ, compressible, switchable video streaming to driver station, and a custom client in Java for setting parameters.