Hello all,
tl;dr:
I need dataset ( idealy 10 - 30 (for now, as start) ) pairs of timelapse videos with the G-code that was printed. Preferably
failed prints. If you have any that you can share with me / send to me I will be thankfull.
Preferred requirements:
no less than 480p, idealy 720p and up
the larger print the better, maybe 4h and up
More details:
It's for my semester project for computer vision class (CS Engineering). The idea is, to be able to recognize difference between rendered G-code and object being printed to detect printing fail. Not real time, just every layer or so, enough to save possily wasted material and damage to your printer and to not be too demanding on computing power required.
Its supper early to promise anything from this so please dont get your hopes up yet,
In the worst case - your files will be archived and forgoten with me.
Best case - plugin for Octoprint that could detect if the print is going OK or if something is wrong (detached from bed, excess stringing ...) and act on it ( send email notif or pause/stop the print )
Even if I fail I will share my code, so other can lear from it.
I already have some timelapses of my own, but i need more from different enviroments, backrounds, lighting and so on.
And i really dont want to do simulated fails on large prints if i dont have to.
Thanks for any response and help.