11-29-2011 09:04 PM
Hi all,
I am working on a project for a class and have been for quite some time but continually hit roadblocks. Right now, I am trying to locate an object that I place over a series of horizontal black lines I put on the floor. I would like to be able to find roughly the center of the object on the floor. My current approach is to use the line profile tool and have it "step" across the image in 5 pixel groups. I do this for a "blocked" image as well as an unblocked template. My real trouble is interpreting this data into something useful! I'm sure there must be an easier way to go about this kind of problem but I just can't think of one. I have attached my test VI as well as a few of the images I am using (I hope that I made that folder correctly). Thank you in advance for any help you can offer!
-- Kevin Lakawicz
11-30-2011 09:29 AM
I should note that the above file is a work file and is probably what you might call incredibly messy. If what I am trying to do is confusing, feel free to ask. Also, I have recently been looking into image subtraction as a possible alternative. Is this a viable solution to the problem I have? I tried to modify the example subtraction VI but it kept giving me an error about the size of my images. Again, any help is appreciated!
-Kevin
11-30-2011 04:44 PM
It looks like from the test images that you should be able to do a perspective calibration based on the distance between the lines. This would then allow you to do pattern recognition on the object. This would only work if you were searching for the same object each time, but if you were it might be a more straightforward approach. Here's a link to the perspective calibration example. Try it out and see if this approach works better for you.
Perspective Calibration Example
http://zone.ni.com/devzone/cda/epd/p/id/5631