Add Camera Benchmark Tool and Allow Correct Unprojection of distance_to_camera depth image

[glvov-bdai]

Description

This change allows a user to easily test different camera settings/configurations to see what works best for them.

This also includes a fix to be able to correctly unproject point clouds from the distance_to_camera replicator,
as the unproject_depth method assumes that data originates from distance_to_image_plane replicator and as such could previously result in distortion when previously used with the distance_to_camera replicator.
pyproject.toml

Type of change

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• This change requires a documentation update

Screenshots

Checklist

• I have run the pre-commit checks with ./isaaclab.sh --format
• I have made corresponding changes to the documentation
• My changes generate no new warnings
• I have added tests that prove my fix is effective or that my feature works
• I have updated the changelog and the corresponding version in the extension's config/extension.toml file
• I have added my name to the CONTRIBUTORS.md or my name already exists there

[jsmith-bdai]

LGTM with a few minor comments. Thanks for adding this useful tool 🎉

[pascal-roth]

Thanks for making this tool!

I would have a proposal for the general usefulness. I find it hard to estimate how many cameras I can use in my actual scene, and I basically have to go back to the same trial and error as before, even if I used the benchmark. Personally, I would expect that I pass my scene definition (including robots, terrain, and other sensors), plus the intended data types, and then the script would add the different cameras, give me the memory consumption of each one of them, and tell me how much of each kind I can use.

[glvov-bdai]

@Mayankm96 @pascal-roth

Manager based vision cartpole MR at #995 uses same convert_perspective_depth_to_orthogonal_depth utility