How to get, for a same sequence and a high fps, the color video and the depth video which is mapped to the color one? RRS feed

  • Question

  • Hi all,

    I would like to know if it is possible to record, for a same sequence:

    - the color video

    - the depth video which is mapped to the color video.

    In fact, I want to know the depth in some pixels of color images, but, since my object of interest is deformed during the sequence, I want to know if some codes exist.

    Then, is this following idea possible:

    Recording the color and depth video with the Kinect Studio, then performing this Coordinate Mapper?

    If it is, how can I extract color and depth video from the Kinect Studio files?

    Thanks in advance,

    Monday, June 29, 2015 12:20 PM

All replies

  • The Kinect sends 'frames', if you set it to capture the depth and color frames, that should be what you want.

    You can just create a copy of the frames before modifying them to give you the different variations you want.

    Kinect Fusion in the SDK is a great source for code if you need somewhere to start.

    The coordinate mapper just uses the table located here "this.sensor.CoordinateMapper.GetDepthFrameToCameraSpaceTable()", you can export that table and do the mapping at any point.

    To my understanding, the color/depth is just a stream of images and not in a 'video' format that you could view with media player.  You'll either have to code it to save as a video or use screen recording software.

    Hope that helps some.

    Tuesday, June 30, 2015 5:19 PM
  • Hi Jacob the Geek,

    Thank you for your answer and sorry for the delay .

    I should be more precise in my question :) I know that the Kinect send frames, but I want the higher fps that Kinect proposes (30 fps, I think). I want to know which solution gives the higher fps (a C/C++/C# or getting frames from the .xef file).

    Thank you for the CoordinateMapper function: I will take a look!

    Monday, July 6, 2015 7:16 AM
  • The v2 sensor always runs at 30fps, but the color stream can switch back to 15fps long exposure mode when it detects low light levels.

    Unfortunately there are no controls for that in the SDK so all you can do is add more lights in the room to lighten up the darkest pixels in the image.

    Note that this is assuming your software is efficient enough to also process at 30fps of course, the programming language or live vs .xef should matter really.


    Monday, July 6, 2015 7:50 AM
  • Hi Brekel,

    I didn't know about this long exposure problem. Thank you for the advice!

    I am more confident with the .xef file :) Do you have a good resource to extract streams (and frames so) from the .xef file?

    Thanks in advance!

    Monday, July 6, 2015 9:20 AM