09-12-2014 02:04 AM - edited 09-12-2014 02:05 AM
Hello,
I have programmed a machine vision system for a scientific test (simple blob detection), using one GigE Camera and a DAQ system (USB6001). Working great! But now I want to pimp the project. My target is to use 3 Cameras in parallel, which perform the same analysis for three experiments. I programmed the basic-one-camera-code by using a state-machine, a single loop for the DAQ and using two queues for communication between both loops.
Now my question: Is it wise to extend the code with 2 extra parallel state-machines for the new cameras? Should I integrate all three cameras in one state-machine? (It is not necessary that the analysis for all three cams is done simultaneously! Step by step is okay! )
Paul
09-12-2014 03:11 AM
I would prefer going by a single state machine to handle all the three cameras (Or even more incase if it comes). If you have additional 3 more cameras do you go for additional 3 loops? No. You may add different states for handling different cameras. If you increase the number of loops you are increasing the code and the complexity.
09-12-2014 03:51 AM
If you are able to enclose the whole functonality in one VI then I would just add a higher level VI that launch it 3 times in parallel (show front panel when opened). Of course you will have to provide the configuration as an input but it is then easy to scale it. Another solution would be dynamically loading a number of your VIs.
10-24-2014 03:02 AM
Hello,
I want to thank you both for your tips! In the end I choose to call multiple Sub-Application connected by shared variables. Theoretically I can use now multiple computers. Works fine!
Bets regards,
Paul