Hi HN, At work, I was taught how to calibrate cameras using custom scripts, but, while it worked, the process always felt like a black box. I would run a script, wait, and if the calibration failed or ended up with a massive reprojection error, I was often left clueless as to why it happened. Low contrast on a few frames? Poor field-of-view coverage? I built Visical as a personal project, a GUI tool built with C++ and OpenCV.

The idea behind is to have quality assessment into the detection phase, giving real-time feedback before calibration.

It supports loading images from disk or capturing them with webcams via OpenCV or GenICam cameras via Aravis.

Since this is my first really big solo project, it currently has some limitations: it is currently scoped to single camera setups and supports only chessboard and circles grid patterns. I hope to add support for other camera setups board pattern and other features in the future.

I'd love to hear your thoughts about it!