Forums > Jitter

motion tracking question (newbie)

January 22, 2007 | 10:16 pm

Hi there
I’m working on an installation which requires motiontracking. I thought optical tracking would be the cheapest way to go, so i looked into some patches like cyclops and cv.jit
I got the cv.jit.track object working, but there some things i want to add in terms of functionality
but i’m stuck. Mostly becease i’m self-taught…

anyway; this is the data i want to extract (excuse my poor technical english)

-the number of people in the space
(every person who enters the space should obtain his/her own trackingpoint (id?). I thought i could fix this by placing a sensor at the entrance (like a floor-panel) when triggered it initiates a trackpoint on that given x-y coordinate (0,0))

-so… Every person should be individualy tracked from the intialisation at the entrance

-when you move closer to another person the ‘blobs’ will sort of melt together, so trackingpoints could be lost or ‘transfered’ to another person while moving through the room.. That’s ok, and actually emphisizes the dynamic nature of (any) energy..

this is a link to a scematic overview of my idea, hope this explanation is not too abstract..

so the variables i need:

-number of people
-assigned id-numbers
-x and y position per id-number

does anyone know a way, or maybe a patch which is already suited to do these things? I hope someone can push me to the right direction, cause this slow progress is making me mad..

thanks in advance,
Mike

[img]index.php?t=getfile&id=441&private=0[/img]


January 23, 2007 | 5:40 am

hi, try using cv.jit.label with cv.jit.sort to keep your labels consistante.
go over the list for more info and advice, serach for "cv.jit" "infra"
"tracking" blobs" etc.

On 1/23/07, Mike wrote:
>
> Hi there
> I’m working on an installation which requires motiontracking. I thought
> optical tracking would be the cheapest way to go, so i looked into some
> patches like cyclops and cv.jit
> I got the cv.jit.track object working, but there some things i want to add
> in terms of functionality
> but i’m stuck. Mostly becease i’m self-taught…
>
> anyway; this is the data i want to extract (excuse my poor technical
> english)
>
> -the number of people in the space
> (every person who enters the space should obtain his/her own trackingpoint
> (id?). I thought i could fix this by placing a sensor at the entrance (like
> a floor-panel) when triggered it initiates a trackpoint on that given x-y
> coordinate (0,0))
>
> -so… Every person should be individualy tracked from the intialisation
> at the entrance
>
> -when you move closer to another person the ‘blobs’ will sort of melt
> together, so trackingpoints could be lost or ‘transfered’ to another person
> while moving through the room.. That’s ok, and actually emphisizes the
> dynamic nature of (any) energy..
>
> this is a link to a scematic overview of my idea, hope this explanation is
> not too abstract..
>
>
> so the variables i need:
>
> -number of people
> -assigned id-numbers
> -x and y position per id-number
>
> does anyone know a way, or maybe a patch which is already suited to do
> these things? I hope someone can push me to the right direction, cause this
> slow progress is making me mad..
>
> thanks in advance,
> Mike
>
> [img]index.php?t=getfile&id=441&private=0[/img]
>
>
>
>
>
>


Viewing 2 posts - 1 through 2 (of 2 total)