MobilityGen depth info output

[Marsak24]

Does MobilityGen output:
Ready 3D point clouds (XYZ coordinates), or
Only depth images that require further projection/unprojection to get 3D points?
I ran H1 with the simple warehouse env, I got a couple of files exported but they weren't including depth data at all.
Thank you

[jaybdub]

Hi @Marsak24 ,

Thanks for reaching out!

Currently, we only record depth images.

The sensor modality recording is done on the Camera class.

MobilityGen/exts/omni.ext.mobility_gen/omni/ext/mobility_gen/sensors.py

Line 56 in 23456db

self.rgb_image = Buffer(tags=["rgb"])

However, replicator does seem to support point clouds, it may be possible to integrate this support into the class.

https://docs.omniverse.nvidia.com/extensions/latest/ext_replicator/annotators_details.html

Please let me know if this helps or you have any questions.

Best,
John

[Marsak24]

Dear John,

Thank you so much for your reply.

ok, so I will have depth images and I need to extract point clouds and build the map for my application?

Regards,
Marwa