(Sorry for the machine translation).
Hello. Subject subject following. I want to create a free program for tracking the 3D printing process by the neural network. The task of the neural network is elementary, and I hope to find help in the relevant sections of the NVidia community. The neural network will only have to select the printer 3D carriage and guides in the image, with the printer carriage, and remove them from the subsequent processing.
I will describe a message process:
- print simulation is created, and several slides are made on it by layers, or slow video, with image change on layers. The G-code is modified for subsequent synchronization - but that is another industry, and the modifiers I write myself - is not a problem.
- Further, printing is started with this video, and the neural network only compares the current image obtained from the camera with the corresponding frame from the simulation. The simulation image is changed by command from G-code. The neural network removes the carriage , guides from the real camera, corrects the lighting , and it is possible to compare images at least with it, at least with an algorithm.
What you need in this section - I need to know how to perform logical operations on the models on the table (truncated with a plane from above) in the PhysX. The engine is going to use Unreal Engine, if of course you can free, as the program does free.
The enthusiast himself, the inventor, have patents and patent inventions, when it is necessary to implement something, then I write a code, and this task I solve in parallel with the setting of video surveillance on the family farm, decided that if there I adjust (need forced - climate changes and delivers problems, they need to be monitored), then mood and for the printer.
I 'm grateful for the help in advance.
If I need to seek help on the subject, of clipping the object, to other forums - please give a link.