Bottom detection algorithms

[LOCEANlloydizard]

Hi all,

I'm opening this issue to follow up on the implementation of bottom detection algorithms. A basic threshold-based method and the Blackwell method (adapted from Echopy) were implemented in PR #1521.

The next step could be to add a function that allows users to manually correct the detected seafloor, ideally with dynamic refresh for visual feedback. We'll also need to optimise the current implementations by translating the NumPy-based logic into xarray and dask operations, to ensure scalability with large datasets.

As discussed in #729, the creation of a bottom mask from existing annotations is handled by echoregions.bottom_mask, and the masking itself is applied using mask.apply_mask.

You can see an example of the application between echopy and echoregions in the notebook example from OSOceanAcoustics/echopype-examples#65.

[LOCEANlloydizard]

We want to add a second bottom detection: can be used to measure bottom roughness/hardness, useful in habitat mapping. Also need to add more test to the functions.