Multiple Target Targeting

[RaxusPrime]

Hey guys! I have a question that I've been wondering since last year when I started programming vision tracking. How do you see two different targets (left or right target specifically) and keep then unique from each other? I've read the usfirst programming guide on the vision programming and I saw that the lab-view version of the vision tracking can determine if the target is left or right. How do you do this? Any help would be very helpful .



By the way I have unbundled the target info array so I'm up to that point.

[faust1706]

I gave this thread some time so someone of more knowledge on the topic of computer vision in labview could help, but....that doesn't seem to be the case.

You could find the centers of the two targets, 1 and 2, and check to see if the x component of 1 < x component of 2. I'm not sure how easily it would be to extract the center from a target, but...that is the way to do it. Then, when you know which one has the lower x component, you could save off that variable as left and the other as right.

[Greg McKaskle]

The LV code uses the shape, or rather the relative location of the vertical and horizontal elements to identify a hot left, hot right, or not hot target. In retrospect, that would have been a better way to encode the results.

If you look at the ranking code, you will see it call Rank HV combos. This is where it looks through the list of particles, pairs them up, and classifies them as left or right.

Greg McKaskle

[RaxusPrime]

Originally Posted by Greg McKaskle

The LV code uses the shape, or rather the relative location of the vertical and horizontal elements to identify a hot left, hot right, or not hot target. In retrospect, that would have been a better way to encode the results.

If you look at the ranking code, you will see it call Rank HV combos. This is where it looks through the list of particles, pairs them up, and classifies them as left or right.

Greg McKaskle

Alright cool I think I see where you're coming from, the HV Combo is in one of the sub-VI's right?

Originally Posted by faust1706

I gave this thread some time so someone of more knowledge on the topic of computer vision in labview could help, but....that doesn't seem to be the case.

You could find the centers of the two targets, 1 and 2, and check to see if the x component of 1 < x component of 2. I'm not sure how easily it would be to extract the center from a target, but...that is the way to do it. Then, when you know which one has the lower x component, you could save off that variable as left and the other as right.

Well I really appreciate the help, I didn't think of putting it like that. And I know that the center variable is one of the variables that is unbundled from target info, I'll look into it. Thanks again for your input!

[Greg McKaskle]

I think the subVI is called Rank HV Combos.

Greg McKaskle