Capture joint position on X,Y,Z axes with Kinect V2

Asked

Viewed 316 times

6

I need to extract the position of the joints from a recording and/or in real time, and convert to database format and/or file (txt). Similar to what is done in this video, but instead of displaying need to save the values of the X,Y,Z axes for each of the 26 joints.

In the documentation provided by microsoft is given shows how to track in real time, but not how to extract the raw data from these variables.

Body tracking 1

Body tracking 2

If you can’t save, how can I extract this data directly from the standard format of the files recorded by Kinect Studio (formats .xrf and .xef)?

I also want to know the difference between the classes BodyIndexFrame, BodyFrame and Body

The script to get this data can be done in any language, but I imagine it will be done in C++/C# as the operation of everything else.

c# c++ image kinect
  • 1

    Um... it seems to me you missed reading the documentation carefully. I don’t know much about Kinect, but the first link you mention uses the structure CameraSpacePoint...

  • 1

    Actually, I was looking for a class that did that, not directly by the structure, thank you very much.

  • 1

    Hi again. If you can post an answer with your conclusion/solution, that would be great. So your question no longer appears unanswered and can more easily help other people in the future (besides you get a reputation for it). : ) I thought about posting one, but I don’t have the environment to test a minimum example that works for what you ask.

  • 1

    I’m building the code, so as soon as I can do this accurately, I can post it. I thought about this not only because I completed the answer, but because the Brazilian community that develops for Kinect is not very large and/ or participatory ...

  • 1

    Properly detailed solution.

  • 1

    Very cool! Won my +1. :)

Show 1 more comment

1 answer

2


  • Unable to extract file position .xrf and .xef.

  • Body is the class containing the state and position (x, y and z axes) of all joints of a person A being screened.

  • Bodyframe is the class used to store the output stream of the Kinect V2 (raw data) of each body being tracked.

  • Bodyindexframe is the class that makes the casting of Bodyframe for the data interpretation and display in standard output.

  • The position of the joints are in Body.Joints.Values.Position. [X, Y or Z], example here.

Follows solution.

https://github.com/inessadl/Kinect2Arff

It is an implementation from another, referenced in the project’s own README. The project will receive various functionalities in the short to medium term.

Browser other questions tagged c# c++ image kinect

You are not signed in. Login or sign up in order to post.