DeepStream the X,Y,Z of the detected object

How can i get the x,y,z location of the detected object ?
Is that possible using DeepStream?

Hi
Here is deepstream application architecture, you also can get from here, https://docs.nvidia.com/metropolis/deepstream/dev-guide/index.html#page/DeepStream%20Development%20Guide%2Fdeepstream_app_architecture.html%23, we have metadata associated with each frame, by iterating metadata you can get detetected objects coordinates. https://docs.nvidia.com/metropolis/deepstream/plugin-manual/index.html#page/DeepStream_Plugin_Manual%2Fdeepstream_plugin_metadata.03.2.html

A screenshot of a cell phone Description automatically generated

Hello amycao
I can get the output of DeepStream the detected object location ,bounding box in this form (Top,Bottom,Left,Right) . IS there a way to have the 3d location of the object on space with respect to camera coordinate frame like (x,y,z) or not
Thanks

I think you’ll have map the screen coordinates to real world. XY plane is probably not too hard if the ground is flat. Z could be considerably harder.

I think Nvidia has a smart parking demo app you could look at. I haven’t looked at it myself but there’s probably some code in there you can reuse.

I just had a look. Seems there’s a tracker written in Python you could re-implement in whatever. It can get you the XY coordinates of cars. You can probably adapt it to whatever you want to do.