0

I'm fairly new to programming and I've been presented with fairly daunting assignment at work. I need to build a program from scratch in order to take advantage of the Kinect's motion tracking capabilities to interface with another application.

Some context:

Someone else I work with has built the test program- a console app using OpenGL. The test program consists of a cube inside of a skymap. The camera looks at the small cube, and can be rotated around the cube to view it from different perspectives

Someone else was able to use the sample codes in the developer toolkit to control the test program. The test program now works with motion tracking (swiping your hand to the side rotates the cube; moving your head side to side changes the camera angle so it looks like you are looking around a floating 3D object; walking forward or backward zooms the camera). It works as it is, but...

The problem is this: Now that we know it all works, it's time to simplify everything so that we can run the test program on a tablet. So the code needs to be stripped down to the bare bones. We need to remove everything from the SkeletalViewer code except for the elements that gather and process the data, so that it can be used in another program.

I've been asked to build a console app from scratch (rather than tearing apart the sample code-as this is extremely messy) that allows us to use the Kinect with our test program.

I've spent the last few weeks trying to figure out the code and I'm feeling overwhelmed! I don't know where to start.

Here's my question: what are the absolute bare essential building blocks in the Kinect program? I do not need it to draw ANYTHING. I just need a console app that, when running, gathers the motion tracking data and sends it to the other program.

I would greatly appreciate any guidance you can provide.

Thank you in advance!!!

-JD

user1639742
  • 11
  • Are you working with the Kinect SDK? – EdgarT Sep 01 '12 at 09:06
  • Yes I am. I'm attempting to understand it well enough to build my own app, but I'm still not very familiar with programming and C++ so I don't know what 90%of it means. – user1639742 Sep 04 '12 at 14:39

1 Answers1

0

You don't need to draw anything, but there's a need to create an event for caching the frames in order to work one them.

Here's pretty cool description of skeletal joints: MSDN - Kinect
Another thing I can give you is the main page of Kinect project. Here are the libraries, guides, code samples etc. You can download and install Kinect Toolkit, there are several programs inside (binaries + code samples) - everything you need to learn Kinect API: Kinect For Windows - Downloads and Kinect For Windows - Learn

Nickon
  • 9,652
  • 12
  • 64
  • 119
  • Thanks Nickon. So I know I need to initialize the Kinect, then the next step is to create an event? Do I only need to create an event for the skeletal data? I believe the next steps are to open the stream and process the data, correct? Sorry for all the simple questions. I'm just so new to programming. – user1639742 Sep 04 '12 at 14:44
  • No problem:) Everyone was a rookie:P Yes, the next step is to create an event for skeleton data. You don't need to create events for depth data and image data. If you have more questions, you can mail me nick0n8 (at) gmail.com – Nickon Sep 05 '12 at 07:23
  • Thanks Nickon! I appreciate all the help! – user1639742 Sep 06 '12 at 16:07