Me and my team have successfully implemented the image tracking code, and it is consistent, however it is a bit delayed. We were thinking that the reaction time might be sped up if we had Labview recognize a template rather than recognizing pixel colors. We uploaded the image via ftp to the robot, and we used FRC Vision Assistant and converted it into Labview code. Our problem is that we get an unidentified error coming out from the IMAQ Template Recognizer and we do not know how to even begin debugging the automatically generated code. Does anyone have any suggestions or code which recognizes a template?
I do not think this will speed things up. Templates are quite computationally expensive. Color histogram templates are somewhat less expensive, but the threshold and particle approach in the examples showed to be way faster.
If you want to step through the autogenerated code, it isn’t that hard. Start it running and start opening panels looking to see if the code had a runtime error. If the panel doesn’t update at all, it is because it is reentrant, but otherwise you can track things down a layer at a time.
Greg McKaskle