Merging two depthimages > blobtrack

General TouchDesigner Discussion
1

I have two images coming from 2 kinects, both a few meters apart and pointing down at a stage. Now I want to efficiently merge the two depth images to form 1 big image, which i then can feed into the blobtracker (to get the same IDs across multiple cameras). OpenNi seems to have a projective to real word function, which I assume gives real world coordinates as R=X, G=Y, B = Z. But I’m not sure how to combine these encoded coordinates from 2 kinects so that the result can be blobtracked. Any ideas?

Thanks

2

Hello Achim,

Sorry, I won’t be able to help you.

But I have a question : How did you manage to get 2 kinect’s streams into TD ?

Thanks.

Xavier

3

@xavier: I’m currently trying to update my kinectToTouchDesigner bridge to support multiple kinects

4

deleted prev question, as it turns out z it’s actually the distance from the XY plance and not from the sensor

5

Achim, do you mean that the z value is the distance from the closest limit where the kinect seems to cut off?

rod

6

I am trying to run the KinectToTouch setup with the latest unstable driver versions:
NITE-Win32-1.4.1.2,
OpenNI-Win32-1.3.2.3,
SensorKinect-Win-OpenSource32-5.0.3.4
OS: w7 64bit
The drivers are working well with other applications, but unfortunately TouchDesigner shows no pictures when open KinectToTouch.tox and KinectToTouchDesigner.exe. The SamplesConfig.xml is in the same folder. I am totally new to TouchDesigner, so maybe I missed something.

Thanks for help

7

touch unfortunately requires a commercial license for this bridge to work (it uses shared memory and that does not work in the free version). Do you have a commercial license?

8

Understand. Thanks for the info Achim.
Just looking for an easy way to play around with the depth and pose detection data from Kinect without much coding, using it for displacing a 3d object, attaching particles to the mesh, etc. Like one would do it in Virtools or a 3d render app.