Refactor the vision sensor API.

[Tropix126]

What's the motivation for this feature?

Many vision sensor features are not supported by the current API or are outright broken. For example, you can check for detected objects, but can't register any signatures on the sensor to detect such objects.

Describe the solution you'd like

Hmmm. So, i'm not exactly sure what the API surface for this sensor should ideally look like, I just know it's in need of a refactor. At minimum I know we want a few things:

• Full compatibility with the PROS C++ vision sensor api, except for methods that concern storing signatures on the sensor's volatile memory, since these methods act unexpectedly and are mostly useles for competition.
• Support for storing vision signatures and codes.

Maybe could look at other computer vision crates for inspiration here?

Describe the drawbacks, if any

This would require a breaking change to the vision sensor API.

Describe the alternative solutions, if any

N/A. Extending the current API is pretty much required for sensor use to be useful under the pros crate.

Additional context

https://pros.cs.purdue.edu/v5/pros-4/classpros_1_1v5_1_1Vision.html
https://api.vexcode.cloud/v5/class/classvex_1_1vision