no it would be even easier than doing everything 100% cgi.
the way its done is you have a model / effect for the orb, you let it track the plane, you delete the non visible orb out manually frame by frame
i think this can be done within a couple hours if experienced
I mean, obtaining the camera would take more work than me just doing it in after effects.
But if it's an "elaborate" hoax, then the work would be doable I guess.
Question, how easily is it to match the 2 footages the way they are in this case that one is made with models?