Microsoft's First Chip Brings Tank-Finding Design to Xbox

They seem to be quite dedicated to ToF, as they bought not one but two companies (3DV Systems, Canesta) just to have their own inhouse ToF system [1].

I found two images from the inside of Kinect One, if someone is interested: [2] [3], from the Wired article [4]

The processing system on the other hand might be in-house, or probably with the help of their amazing MS Research team. The demo in this video is absolutely state-of-the-art tracking: [5].

[1] http://en.wikipedia.org/wiki/Time-of-flight_camera [2] http://www.wired.com/images_blogs/gadgetlab/2013/05/20130514... [3] http://www.wired.com/images_blogs/gadgetlab/2013/05/20130514... [4] http://www.wired.com/gadgetlab/2013/05/xbox-one-development-... [5] http://www.wired.com/gadgetlab/2013/05/xbox-one/

This does not sound right. I would genuinely like to see a source where they use time of flight, a range finding technique, to detect tanks at long range in daylight when you could instead use active radar or passive thermal technology (focal plane array/starring array [0][1] , to detect a tank.

(My experience is with infantry-level night vision and thermal scopes. I don't know anything in depth about how munitions detect tanks)

[0] http://en.wikipedia.org/wiki/Staring_array [1] http://en.wikipedia.org/wiki/Thermographic_camera

> The Kinect chips, designed by a team mostly based in Microsoft’s Silicon Valley office, use a technology called Time of Flight

I was under the impression that the (original) Kinect didn't actually use time of flight to get it's depth data. Can anyone comment on this?

The original Kinect did not use time-of-flight technology, but projected a structured infrared light pattern and observed the displacement of the pattern to determine depth information. The Kinect that will ship with the Xbox One will use time-of-flight sensing (probably from the ZCam assets they bought with 3DV Systems).

Which is amazing since it removes the indoor constraint that the structure infrared light pattern approach had.

There are several projects out there that use kinects on robots in order to scan entire rooms, creating 3d maps of them [1][2].

There's also this [3], which isn't entirely related to TOF, but i think it's cool, and it uses a kinect. Sorry for my lapse in topicality.

[1] http://www.youtube.com/watch?v=aiNX-vpDhMo#at=19 [2] http://www.cs.nuim.ie/research/vision/data/rgbd2012/ [3] http://www.youtube.com/watch?v=2CbiOikirrg&feature=player_em...