No, because you need ultra-cold or ultra-low density (or both, as with 21 cm hydrogen) gas. If your mean free path divided by your mean molecular velocity is significantly less than the relaxation time, then the the atom/molecule gets knocked out of the necessary high-energy state well before the transition occurs with sufficient frequency.

With [O III] in particular, it only gets into the necessary state via collisions (that's the easy part) occurring in extremely low density plasma, but then it relaxes via photon emission (that's the hard part). So if it gets knocked around by another collision, then the photon never gets emitted in the first place.