[Noisebridge-discuss] Kinect hacking/mapping?

[Lamont Lucas]

Last time I was in the space, I saw about 3-4 people working with a 
kinect, trying to get depth and RGB frames out to their linux laptop.  
I've recently purchased one with the idea of using it to reconstruct and 
measure a 3D interior, and was hoping to be able to work with people on it.

In short, I have the kinect hooked up to my laptop and can pull and 
record an mpeg like stream of RGB-D data.  But I have no idea where to 
go from there.  I originally planned on mounting it on a roomba and 
combining the roomba's odometry with the sensor output to reconstruct 
the rooms.  But then I saw this:

http://www.kinecthacks.com/kinect-accurate-scene-reconstruction-rgb-demo-v0-5-0/

Which is neater, as it's just a guy standing around waving the kinect by 
hand, and his software is joining like surfaces in order to make a 3d 
model.  But there's more in line with what I originally intended, with 
code, from these guys:

http://ic2020.ca/

I was curious if people were still working with the kinect, and there 
was a good time to meet and help me play catchup.  I'm going to play 
with the ROS framework which has a competing free set of USB drivers 
(the NI vs the libfreekinect) and built in visualization framework.