Spatial position from a depth camera image

Hey. Struggling with this one a bit. For various reasons, I’m having to use an L515 sensor at the moment. It works with the Realsense TOP which returns a depth/raw depth image. I’m trying to work out how I can get xyz positions of (for now) a single person in a space from this information. I can get the z position just by analysing the average of the depth map which roughly gives me distance from the sensor.

Given that I don’t have skeletal data, is there any way of also deriving the xy position of the person in the space. Doens’t have to be mega accurate. I’m used to working with Kinect CHOP data so I’m a bit lost with this.

Cheers!

Hi @MMMCT,

this might be a bit weird suggestion, but could you use @tblankensmith’s MediaPipe for this?

cheers
Markus

1 Like