Combining multiple SIG samples

In hopes to improve the consistency of flag detection, we are working on combining multiple samples. One issue with the vision configuration tool is that I does not appear to allow you to combine new and old signatures. Actually, I’m not sure why signatures are sometimes pulled back in and sometimes not. The bottom line is this. Is there repeatable way to capture SIG1 (red flag) at the start of the tournament then capture SIG 2 at a later if lighting conditions changed such that either would be identified. I think we have a handle on how to code the logic. However, because the hardware configurations are obscured in VCS, it makes additive vision configuration challenging.

I suppose one way you could do that is by adjusting the brightness setting on the vision sensor. You could have a look at the object count at the start of the match, and increase the brightness count until you see the expected number. It’s not setting the signiture to be additive as you’re saying, but I believe it would have the desired effect. I’m currently looking into implimenting something based off of this method.

Also, assuming you were only looking for one color of signiture in your program (either red or blue, but never both), you can just configure red as signiture one, and different lighting as signiture 2, and use vision.takesnapshot(0) to gather a table of all the vision sensor data for both signitures 1 and 2. You might have some repeated data, but it could work for what you want.

I used the export utility to export robot-config.h and recreated the program in VCS pro which exposes the SIG codes. From PRO its easier to copy the codes out and back in if needed. It would be really nice if you had the option to save the changes back(or not) to VCS on closing the vision configuration utility. I’m sure that there will be teams that unknowingly clear their signature values when it closes.

It’s hard to see green from a robot’s perspective.