If something changes in whether or not it has an extrinsic (relational) property this does not count as a qualitative change; it is sometimes called a Cambridge change.
So according to Leibniz's law, it is not possible for a banana to both have and not have the property of being green, so it cannot change from being green to not being green.
We can see that Leibniz's law allows this by making explicit its implicit reference to time:
We seem to be able to make sense of this second way of talking. Perhaps it requires thinking of ourselves as commentating on the world from a point of view outside space and time.
(Lowe claims that we use tenseless language when making mathematical claims. Is he right?)
Here is an argument that if presentism is true then there is no qualitative change (and so nothing to be explained): If presentism is true, then the only sentences of the form ‘x has p at t’ or ‘x does not have p at t’ which are true are those for which ‘t’ refers to the present time. So the only sentences of the form ‘x has p at t but x does not have p at t'’ which are true are those for which both ‘t’ and ‘t'’ refer to the present time. But no such sentence can be true. So it is never true to say ‘x has p at t but x does not have p at t'’. So there is no qualitative change. What should we make of this?