Flux calibrating these spectra even in relative flux is not trivial. Yes the standard method to correct for instrument response and extinction using a reference star measurement is ok in principle but there are a few extra things to watch out for which make it tricky when using this technique with wide field moving targets like this. Specifically flat fielding, background subtraction and differential extinction.
Flat field correction of any slitless spectra is complex as each point in the field is a combination of zero orders and diffracted light from other points in the field so there is no one to one correspondence like in conventional flats. In practise this is effectively impossible to untangle. With static targets you can get round this problem by measuring the reference and target spectra in the same position in the field, which is the way I recommend using the Star Analyser for example but this is obviously not possible for meteors of course so my suggestion would be to take a series of spectra of a bright standard star at different locations in the field and asses exactly how much effect it has on the spectrum. If the spectrum shape varies significantly then some allowance has to be made for this depending on the location of the meteor spectrum.
Similarly sky background subtraction is difficult compared with narrow field fixed targets where the sky can be measured directly above and below the spectrum. Perhaps this is not too much of a problem for short exposure videos of meteors though where subtraction of frames before and after the meteor could be used. (Linearity of the light response is obviously important – no gamma correction to be used)
Extinction effects over such a wide field can be significant and will vary along the trail. These could be corrected for using an atmospheric model and some sort of mean elevation figure for the meteor though.
Robin