XML Feeds

.

[jitter] Re: Re: cv.jit tracking - LED's

Stuart Binns spbinns at hotmail.com
Thu Jan 3 13:30:17 MST 2008


Thanks for all your help - I really do appreciate it.

I have tried the 'sort' object, but it causes a problem for the type of installation I'm making.   The two blobs (formed from the LEDs) are mounted in a small handheld box (similar in size to a mobile phone), and the user moves this box whilst pressing on a stretched piece of dark canvas. The blobs appear on the underside of the canvas as white blobs (I used dark canvas so my camera would only see the blobs and no other light soures), which my iSight camera is facing. It is the blobs on the underside of this canvas that the patch follows.  The trouble with the 'sort' object is that when it looses an object/blob, I may assign it a new value when it reappears, so my tracking then fails as I cannot extract XY data from values 1 and 2 for example, as they have changed to 3 and 4. This was brought up in the forum recently, but unfortunately there was no further information...

http://www.cycling74.com/forums/index.php?t=msg&goto=118099&rid=5914&S=2dba38dbb2631b21f0239d019bc0b7c8#msg_118099

I tried the above, whereby banging the 'mode $1' on the cv.jit.label object until the labels correct themselves, but does not work very efficiently.

Ideally, I would just like to be able to track both blobs, but although I selected '@mode 1' on the cv.jit.label, it does not set value 1 to be the biggest blob (as i thought it would), but the blob that appears closest to the top left corner. This was brought up here...

http://www.cycling74.com/forums/index.php?t=msg&goto=120322&rid=0&srch=cv.jit.blobs+erik#msg_120322

It seems that the 'sort' object is the object I need, but then there are it's problems- Catch22!!!

I hope this all makes sense...

Your help would be greatly appreciated.


More information about the jitter mailing list