I was wondering if any other teams are having a problem with the vision sensor. Its lenses seems to be faulty and does not identify colors the way I believe it should. Mechanically, my team and I have tried to solve the issues by cleaning glass and checking wires. Now, I am just wondering what could be going wrong electronically. It it a problem in my code? I thought this sensor was already set up and easily integrated using VCS but we cannot seem to get it to work. Its like its a foreign, 3rd party object. Can someone please tell me why my robot (with the vision sensor) has an odd tendency to launch balls at my fellow teammates and by-standards autonomously? (keep in mind this is my fist year and I studied agriculture up until I joined robotics)
The robot uprising has begun. Nobody’s life is safe. Run while you can.
I can’t help out too much unless you are able to provide us with important things such as the vision sensor’s interface, especially when running the code/setting signatures. Try figuring out where the problem lies - is it the code or the setup of the vision sensor itself?
(I’m assuming you’re trying to aim at flags) When you’re setting your signatures, try taking a signature of the flag, and then see if it can recognize other flags without picking up on visual noise too much. This may take a few tries and some tinkering with the tolerance values. Also, my team found that lighting is super important when setting these signatures; I believe we had to cast a bit of a shadow on the flag to create a gradient when setting the signature.
If that’s done, then the next step would be to look at the code - is everything correct there? Something that I suggest trying is “importing” what the vision sensor sees onto the brain LCD. When the vision sensor detects an object, it returns a rectangle with coordinates and width (iirc) which can be used in programming. Using these coordinates, you can draw rectangles on the brain LCD, which can be useful when debugging.