Consider refactoring core acquisition package

[glopesdev]

The Aeon.Acquisition package has grown overtime to include multiple computer vision operators which bring with them specific dependencies, e.g. SLEAP for pose estimation, Pylon and Spinnaker for camera video acquisition, etc.

This makes it harder to adopt the Aeon framework for simpler experiments which may not require all of these specific components.

The idea would be to split the acquisition package into the following three packages:

• Aeon.Video: all video capture interfaces and video controller / triggering devices
• Aeon.Vision: computer vision tasks such as ROI activity, blob detection, etc.
• Aeon.Vision.Sleap: specialized package for SLEAP pose estimation (to isolate tensorflow and CUDA dependencies)