Hello,
I need to extend the area so…
I would like ask you what is the best way to have the merge of two kinects for the recognition of one skeleton in the area?
Cons (spread)? some ideas/suggests?
Thanks in advances.
hei martin, i’m not sure if this is actually possible. to my understanding you cannot combine the input of two kinects and then ask for the skeleton tracking to be done on that joint pointcloud… have you seen this somewhere?
I think that you need to place the 2 kinects, transform each skeleton in order to align them on the same place add them and divided by 2,
if one skeleton is not present, use avoidnill and input the remaining skeleton on the default pin,
I did not test this, not sure if will work, but…
@joreg I don’t seen this but my idea was to combine 2 Kinects together in such a way as to obtain 1 single wide-range.
In alternative what i did to solve at this my issue I just took their range of each camera and used it individually but with the possibility that the two kinect will change the projection contents together.
Thanks. The issue is when there is some more of 1 skeleton in the area. Anway it is not really clean so I preferred to take another way!
Another problem:
Do you know why with one of my computers, the new one, (i7, 8RAM, 2,6 GHz, Nvidia 4GB, Windows10) with the Kinect (360): skeleton and the depth nodes the reading is sobbing It is not fluid, constant, but intermittent.
Maybe insufficient USB bandwidth ?
How can I understand it?
Rgb node camera work normally!
Are you suing openCV to track depth? If so there is a bug which makes it stutter every second or so, one frame drops, to avoid that change the format/size of texture that goes in openCV image to be 256 wide not 512 x 424, force it 256x256, it’s an openCV image pack bug, broh. Anyone wanna fix that bug btw, you crazy motherfuckin taqueros?
If you are not using openCV on depth then ignore this comment. good luck
This topic was automatically closed 365 days after the last reply. New replies are no longer allowed.