does anyone have a camera matrix which distorts the RGB image of the kinect2 …so it fits the depth|world image.
this could be usefull for futher kinect2 - projector calibration as well as for simple coloured pointclouds.
Any help would be really appreciated
saw some .fx by vux somewhere recently for k2. can’t find it on git atm.
MapDepthRGB & MapRGBDepth
and this might also be interesting sdk coordinatemapper class
there is a rgbdepth node by sebl which puts out a texture that can be used for undistortion. However this undistortion Texture must be related to a kinectv2 specific camera matrix. … i think…
this rgbdepth is kind of a lookuptable that gives you a rgp pixel coordinate for every depthpixel. internally it uses the coordinatemapper. the same node exists for kinectv1… i guess there must be a helppatch somewhere, but i’m not on my dev-machine right now
nice sebl thanks…my problem is i do believe that the coordinatemapper is using a 4x4 or 3x3 matrix …similar like texture transform to create this look up table…and im searching for it…Microsoft said they wont release this information but pointed out that you can find it by calibrating the two images
some news about this topic? i’m really interested in understand how to align rgb and infrared/player with kinect2 :)
thats what the kinecttoolkitdx11 contribution is made for :)
just connect rgb,depth,rgbdepth node of kinect1/2 to the pointcloud node and there you go.
help patches are included.
thanks tmp, nice and useful contribution ;)