Hello,

I have a mono channel depth image coming from a Kinect (512x424 pixels).

I created a 512x424 grid.

I want to add a custom attribute “kinect” to each point of the grid where “kinect” is the value of the “corresponding pixel” (same row, same column).

I try to use pointwrangle and write something like :
@kinect = @pixel.r;

But I have no idea how :
- connect the COP output as the second input ?
- how to get the pixel value in VEX (I am aware that @pixel.r is “pseucode” VEX) ?

I'd rather avoid uvtexture or something similar in order to avoid any filtering (I need a perfect 1 to 1 mapping between pixels and points).

Thank you for any help or clue possible :-)
Nicolas.

Hi
Try the “Attribute from Map” SOP node. And set “Export attribute” to “kinect”

-b

Hi

Thanks a lot, it looks like it does the job :-)

I feel I just need to be carefull with all the options to avoid any resampling or post process.

Just in case, this is waht I did :

In Color settings :
- Source Color Space = Linear
- Do Remap = false

Filter Settings :
- Filter = Point

if somebody thinks I messed up with the parameters, you're very welcome to let me know.

Thanks again Bonsak,
Nicolas.

Sorry I forgot, is there a way to link the “Texture Map” channel to a COP net output instead of an external file ? Because I did not manage to do it.

I'd like to process the image in Houdini before applying the attribute (to create a threshold for instance)

Thanks,
Nicolas.

You can set “Texture map” to

-b

Thank you Bonsak :-)