While it's not trivial, assessing un-improved land value does not seem too different from other complicated accounting tasks that both governments and private organisations have been performing for a long time. Cost segregation studies are already a thing.

A good starting point can be the relevant part of the Georgist primer published on ACX:

https://www.astralcodexten.com/p/does-georgism-work-part-3-c...