Kinect - Vision Assistant

[gallagher]

Hi all,

I am about to start working on a project with Kinect and Vision Assistan (and later with LabView) and I would appreciate if you could give me some advice on what should I do in order to make Vision Assistant let me acquire images with the Kinect camera.

I have already installed on my computer both "Kinect SDK" and "Kinect Developer Toolkit" (versions 1.7 and 1.7.0 respectively) as well as LabView 2013 and the Vision Development Module, version 2013 too.

I have also installed the correspondent drivers from this link: http://joule.ni.com/nidu/cds/view/p/id/4409/lang/en

Kinect camera is recognised properly by the computer, as it is shown on the Device Manager of the Control Panel. I have even tried the Kinesthesia example VI, and it works alright, so I guess the issue of not being recognised on Vision Assistant has nothing to do with the installation of the camera.

Any help would be really appreciated 🙂

Best regards!

[MoviJOHN]

The kinect does not appear as a regular USB camera, so it cannot be directly accessed via IMAQdx Vision Acquisition.  At least this was my experience in the past.

If your primary goal is to capture the 3D map as an image, you can get a trial version of RoboRealm, and connect to the Kinnect there, and then install their Virtual Camera Driver, which is IMAQdx compatible, and then perform the acquisition on the virtual camera image.

The main downside to this is that you only get an 8-bit RGB image, and you can't really don any of the cool stuff with the data, like skeletons, and such.

[Klemen]

Hello,

you can also use PCL (point cloud library - search google) to build a DLL (dynamic link library), which you can call in Labview. For example:

https://decibel.ni.com/content/blogs/kl3m3n/2013/01/08/using-microsoft-kinect-to-visualize-3d-object...

In this way, you can obtain both RGB and calibrated XYZ data (i.e., point cloud with texture).

Best regards,

K