Yes flux calibration of meteor spectra, even in relative terms is going to be tough. The instrument response can be corrected for using a reference star and provided atmospheric conditions are stable, even the atmospheric extinction con be accounted for by taking reference stars at different elevations and applying a correction based on the elevation of the meteor (The usual simpler technique of chosing a reference close in elevation to the target to cancel atmospheric extinction is problematic here of course !) A big problem though is likely to be flat field effects which can be severe with these sorts of wide field systems. A normal flat field correction does not work with slitless spectrographs. The usual workround is to place the reference star at the same position in the field thus cancelling gross flat field errors but again this is tricky here, particularly if the spectrum is generated by integrating along the track. Measurements of reference stars at the apropriate locations in the field the night after the meteor observations could perhaps be used though. Measuring standard star spectra with the star positioned at different points in the field and comparing them would give an idea of the severity of the flat field problem.