none
Data Structure to store information produced by Kinect RRS feed

  • Question

  • Hi guys,

    I'm totally new to C# and Kinect so i'm having a lot of problems here. To get started, i need to have a data structure to store the .OBJ file created by Kinect. But does anyone know if there is a code for the data structure online? I can't seem to find it.

    Basically, using the Kinect device, after i created a 3D mesh of the scene captured by Kinect, the .OBJ file stores information of vertices, normals as well as face. I'm supposed to extract these data and store them into the data structure so i can extract these information later on.

    Can someone advice me on this? Or if u happen to have a link, mind sharing it with me? >_<

    ps: Thanks for even reading, i'm really new to all these stuff T_T

    • Moved by Mike Feng Monday, May 20, 2013 6:58 PM Kinect
    Monday, May 20, 2013 3:15 AM

All replies

  • Look at webpage below.  I belive in som eof the download is source code that will get your started.

    http://www.codeproject.com/Articles/151605/Kinect-and-WPF-Getting-the-raw-and-depth-image-usi


    jdweng

    Monday, May 20, 2013 7:33 AM
  • Thanks for responding! =D Hmm.. But that would require the use of OpenNI?

    I thought of creating a class to store the XYZ vertice, the XYZ normal and the face. So each class object to represent a particular vertex.

    Do u think that would work? What i am not sure about is what does the face data actually mean. I googled and wiki-ed but still i dont really understand. I know that each vertex has a XYZ coordinate and has a surface normal, but i'm not exactly sure what is the face data. For my method to work, if there are 10 vertices, there will also be 10 normals, but will there be 10 faces as well? I'm so confused T_T

    Monday, May 20, 2013 2:50 PM
  • Is the .OBJ a Wavefront file?  ( http://en.wikipedia.org/wiki/Wavefront_.obj_file )  You might have more luck finding software if you search for that.

    Mike


    • Edited by PSU Mike Tuesday, May 21, 2013 12:21 PM
    Tuesday, May 21, 2013 12:19 PM
  • Have a look at the Kinect Fusion Explorer samples. These samples have the capability of exporting a constructed input into obj format. Keep in mind these are starting points as a way to get the data from the sensor and will require that the data will still need to be cleaned up.

    Tuesday, May 21, 2013 6:34 PM
  • Hi Carmine, yeah I finally understood that. Problem now is creating a data structure to store the details. I'm wondering is creating a face class which stores the 3 vertices and 3 vertice normals could work.. Since the number of face is always 3 times lesser than vertices I assume it can work >.<
    Wednesday, May 22, 2013 12:47 AM
  • I referenced the code Project Webpage as a 'talking point', and not necessarily a final solution there is a lot of useful information on that webpage.  I don't recommend using 3rd party API without having source code that I can recompile.  I don't trust executables except from reliable sources.

    I like the following paragraph

    The Power of WPF
    WPF offers great advantages over WinForms considering user experience. Furthermore, WPF's System.Windows.Media is way more powerful than WinForms' System.Drawing. WPF uses WriteableBitmap and ImageSource instead of Bitmap and BitmapData. As a result, I had to rewrite much of the initial code to make it WPF-compliant.

    I don't think that WPF is needed, but do agree that using System.windows.Media is the class that should be used.  Here is another code Project that give a good explaination of how to handle shapes

    http://www.codeproject.com/Articles/129951/Surfaces-in-Windows-Presentation-Foundation

     


    jdweng

    Wednesday, May 22, 2013 2:53 AM