Ray casting-based depth image renderer

[Divelix]

Discussed in #80

This PR adds functionality of depth map extraction from wavemap octree at any pose and camera intrinsics.

I also added python bindings for:

• PinholeCameraProjectorConfig - I use it to store camera intrinsics on python side
• raycast - allows to cast ray between 2 points and get distance to closest occupied leaf
• get_depth - uses raycast for all pixels in depth map (utilizes QueryAccelerator for faster processing)

Moreover, I added corresponding python script to examples folder (I used pathlib instead of os.path because it is considered better practice in python).

[victorreijgwart]

Thank you, @Divelix, for this PR! This feature will be useful for many applications, and I'm looking forward to including it in an upcoming release. Here's the plan for the next two weeks:

• Import the branch from your fork into the main repo so that we can run it through all CI pipelines (similarly to Fix Mypy error #79)
• Move the core functionality into the C++ library to ensure the feature is available in both C++ and Python
• Extend the implementation to support all projection models in wavemap (pinhole cameras, standard LiDARs, and Ouster LiDARs)
• Add unit tests to ensure correctness and so that our CI pipeline can check that there is no undefined behavior

In parallel, we're preparing another PR to extend the QueryAccelerator to work with HashedChunkedWaveletOctrees. We'll complete that first so that this depth renderer can be released with full support for all commonly used map types.