Here's an uneducated guess, based on an evening's googling.
An earlier version of the poem appears in the Carmina Burana, §105 in this book but usually counted §142:
Tempus adest floridum,
Surgunt namque flores,
Vernales mox in omnibus
Iam mutantur mores.
You have to fudge Vernales to fit the melody, which seems strange, but another source supports this. It seems OK if you make Ver- an upbeat.
Translating both literally and metrically:
Now has come the florid time,
For flowers are arising;
Soon the ways in everything
Will be changed to Springtime.
The following four lines continue the idea of evidence that Winter is now in the process of changing to Spring:
Hoc quod frigus læserat,
Cernimus hoc fieri
Per multos colores.
What is injured by the cold,
Heats are now restoring;
We can tell it's happening
By the many colors.
The Biblioteca Augustana's version is probably more authoritative:
Tempus adest floridum, surgunt namque flores
vernales; mox in omnibus immutantur mores.
hoc, quod frigus leserat, reparant calores;
cernimus hoc fieri per multos colores.
The word immutantur agrees better than iam mutantur with the idea of "being changed into" and the transitive use of muto.
The punctuation indicates that the curator thought that vernales modified flores. Leo Tepper posted the same conclusion in his blog. But don't jump on that bandwagon just yet. There's more evidence below.
Wikipedia reports that a "sanitized version" of Tempus Adest Floridum was published in the Piæ Cantiones. No source is given for that characterization, but the next two verses of the Carmina Burana version are about virgines cum clericis finding love in the meadows—probably not very pious by the standards of the Swedish Reformation. The version in the Piæ Cantiones still describes meadows, but now it's only about plant life.
So, Tempus Adest Floridum got an edit, which fixed both the meter and the naughtiness. (The meter is still flawed, since the unstressed syllable Ver- is now on a strong beat, but this minor flaw seems common in Medieval Latin poetry.) The idea still seems to be that we are now witnessing the change to Spring. Per multos colores, in line 8, was revised to per multos labores, reinforcing the idea that something is causing everything to change.
Now, consider the medieval script in which immutantur would have been written. The letters -immu- would likely be a series of vertical strokes. If written inmutantur, the editor of the Piæ Cantiones could easily have mistaken the eight consecutive minims as the start of immitantur and then "corrected" the double m. (Perhaps iam mutantur is another transcription error introduced due to the same ambiguity.)
So that's my hypothesis: imitantur is a transcription error. The actual verb is inmutantur or immutantur, its subject is mores, and its object, properly an adjective, is vernales.
Commentary on p. 253 of this edition, from 1910, reports another likely transcription error in the Piæ Cantiones, in the second verse.
I welcome debunking from others more knowledgeable (or less). Some bases for doubt: palæography, transitive inmuto, and the erudition of Jacobus Finno and the other editors of the Piæ Cantiones, who presumably knew what they were doing.
After some more googling I found the relevant passage at the bottom of folio 58v in the Carmina Burana:
It clearly reads iam mutantur, so now I wonder where the Biblioteca Augustana got immutantur. If that's a period after omnibus, then vernales modifies flores as in Joonas's hypothesis, yielding this literal translation: "…for the Spring flowers are arising soon in everything; already the ways are being changed."
After still more googling, I didn't find mutari used in the transitive way required for this hypothesis, but I did discover the impersonal passive, explained in this answer. The impersonal passive makes sense of both the mutantur and imitatur versions, without needing vernales to (jarringly) modify flores, so I'm judging the transcription-error hypothesis an error.