hi guys! I just started with an installation with 2 kinects. Till now I discover that merging two depth is not that easy, so I guess I’ll go for the pointcluod. Any advise on the better way to do it?
area to cover is 6 mt wide and 4-5 mt depth, shape is in the image. I found that horizontal field of view should be 57°, it’s right?
Don’t need skeletons, just shape of people
in case is gonna be too complicated I’ll go for only one kinect, but little part of the area will we cut out (or at least it seems like this on my preview
there is auto calibration possible using player as calibration object…
i’m not sure this gonna work with two kinnect one, since quality of tracking on K1 is crap, there are multiple ways how to find where kinnect located in the world space.
Easiest prolly is ellitos calibrate projector, however i think i placed few boxes and just aligned kinects
have a look at the calibration node in my dx11.pointcloud pack. there is a help patch included.
if you download the latest version @ https://github.com/letmp/dx11-pointcloud you can also find a girlpower patch how to use two kinects in your setup. it is called “MultiKinectSetup”. You only have to do the calibration for both kinects and you are done. Have fun :)
yes, working with kinect 1 here. you need a different USB host controller for each kinects, looks like normal desktop computer have usually 2 (one for the front usb and one for the back)
also some notebook have also 2… plus, I think you can connect external stuff in order to have more usb host (but never tried)