Kinect2 - Depth Image Mapping

Hey folks,

I’m completely new in vvvv. I want to create a topography map in areas based on the depth sensor of the Kinect 2 and have no idea how to start. I just googled about 2 hours to find something helpful but as a beginner it’s very hard to understand the connection between things here… Does anyone of you has an approach how I could start?
The Kinect already works in vvvv and I get a depth image in my renderer (I guess… it’s just a red/black image?)

Kindly regards

Patrick

2 minutes in the forum search later and

I guess you would want to displace a mash with the depth map…

Thanks… but this isn’t very helpful when you don’t have an idea how to use them with a Kinect; every tutorial is like written for advanced/pros or non-existend (or I use the wrong words)… I just can handle the ultra basics of some animations and the basis “user interface”. I tried it to get work for 2h and I havn’t reached anything :/

I would be very pleased if someone can give me a kind of “roadmap” what steps are nessecery to reach my goal. Not a step-by-step walkthrough; just some bulletpoints that I can google to get closer to my target :)

Here are my ideas how I would start:

  1. Get a depth map from the Depth-Node from the Kinect 2 (already done [the black/red stream])
  2. Make it to a texture/Displace as a mesh (however; no idea; no good tutorial found)
  3. Lay the shader on it (however; no idea; no good tutorial found)
  4. Display it in the renderer (one item I know how to do ^^)

Am I far away and are there some tips?

Hope you guys can help an absolute beginner. Thank you!

Look at the DX11 Pointcloud pack, I believe it has a node to generate a mesh from the Kinect depth image.

Once you have that mesh, then you can feed it into the shader of your choice such as one of the Topo examples. I have a DX11 version of the Topo Phong shader above, I just need to post the update for it (the one above is DX9).

Thank you mediadog! Now I get an RGB Image of myself and a “gray layer over it”. Could you please upload the new shader version? :)

PS: I tried other shaders if something happens… but it still stays gray :/ Any ideas or am I doing something wrong?

You need to hook a Camera to your Renderer so you can look at your scene from different angles to see the 3D shape. I like Camera (Transform Orbit) myself.

I think that “grey blanket” you are seeing is where the mesh does not exist or is flat because the Kinect depth points were outside limits. Try playing with those input parameters to the Mesh node.

And for a shader, try something simple first like Constant. Then if you want to do a topo-map kind of thing, look at ConstantProjection (DX11.Effect).

Ah thank you! I will have a look this night when I’m back home! :)

This topic was automatically closed 365 days after the last reply. New replies are no longer allowed.