Add ARM NEON intrinsics to unpack_yuy2

[fateshelled]

trafficstars

Changes:

• adding unpack_yuy2_neon function
  • unpack yuy2 to y8, y16, rgb8, rgba8, bgr8 and bgra8 format.
  • tested on Ubuntu 22.04, OrangePi5 (RK3588s, 8GB of RAM) and RealSense D435.
• adding early continue in align_images function
• fix rs2_project_point_to_pixel and rs2_deproject_pixel_to_point function
  • change order of operations

[sysrsbuild]

Can one of the admins verify this patch?

[fateshelled]

I explained too little, my apologies. I added the optimization code for ARM CPUs because the CPU load was too large when using ARM CPUs. Please review if you like.

[Nir-Az]

Hi @fateshelled , Thanks for the PR, it will take some time but we will get to reviewing it. Maybe you can undo format changes to make the PR more readable? I see rs.cpp have format changes added..

[fateshelled]

Hi @Nir-Az , Thanks for the reply. I fixed format changes. Please review it.

[Nir-Az]

It will take us some time to review and validate this PR on dedicated HW. I would suggest to split bug fixes and new features as we may merge 1 faster than the other :)

[fateshelled]

Thanks for the reply. I have changed the request to a pull request for new features only.

[Nir-Az]

@fateshelled appreciate your contribution, we are always happy to integrate community pull requests :) Since this feature looks safe for regression and would benefit you and other users, I merged it based on your testing.

Thank you :)

[fateshelled]

Thank you for merging the PR. I am very happy to contribute.