I want to be able to detect persons on a grid of 5x5 meters, what do you guys think? should I use the kinect? of make a setup with cams or sensors?
the idea is to make an installation of 7 guns pointing at anyone within the grid. I did that already: http://www.artvantriest.nl/pelotonfilm.html but I am making a second version and it needs to be more precise… am using car-window-wiper-motors now and am looking for a good way to detect the people in front of my interactive peloton…
if anybody has a clear idea about that, youre very welcome!!
When I was testing it, it could detect a person from about 50cm (distorted skeleton, because you are not completely ‘seen’ by the Kinect, but if you don’t need skeleton it’s ok), up to less than 4m. Distances > 4m are simply not detected.
The ‘usable’ (as in full skeleton tracking) space feels more like 2x2m, maybe 2.5x3m, but not more.
So it depends on what you want to do, but it’s definitely quite limited.
if you put it on the ceiling of a tall room, looking down, using the depthmap, i think you should be able to detect humans in a 5x4mt area… people’s head will be 1,5/2 meters closer to the sensor, so it could work. just my 2 cents
i agree with sapo, use that setup + contour
you will find a problem here, how to extend the kinect cable, i try soldering + utp shield cable and it didn’t work
i need to try with a usb2 hub , if someone have any idea over this please let us know.
