07-22-2011 12:17 PM
So i've been experimenting with creating a force-feedback system for a piezo stage and laser trap. For the most part I have it working; It's a little messy but I've only been teaching myself labview over the past couple weeks. Anyways, I'm looking for tips on how to raise my frame rate. It's pretty important that while running the VI i dont really have more than 2% deviation in the frame rate (so it has to be fairly steady). Today I noticed something that was really eating my frame rate was that I'm currently saving the tracking information to an excel file on every iteration, deleting it almost doubles my frame rate but having the information for reference is pretty important too.
The first method I used for tracking involved using imaq count object 2. and then decreasing the roi by about half for frame rate consideration and tracking that way.
The other method I used involved having the user draw a line, then having the program bin the image to ONLY that line, and using imaq edge tool 3 to find the darkest point on the line and track it. (although this one is far less consistant than count objects 😕 ) Is there another way to find the darkest set of pixels for tracking?
Since writing the pixel data using write to spreadsheet is eating so much processing power, is there a way to do it that isn't as demanding?
Thanks!
EDIT: Underlined actual questions for quicker reading, the rest is background information, although pretty important.
Also my exposure is about the lowest I want it right now. So the tips I'm looking for are in writing the VI cleaner. I have attached both methods I explained above.
07-22-2011 12:22 PM
Something else to mention: I was wondering how event structures would work in this case and if they would save processing power. I was looking at examples VIs that had them trying to decipher how to get them to work. But it seems event structures run off of front panel interaction, and not feedback information. Any pointers?