Observations need to subtract continuum if it is there. Neighboring filters. Beware of neighboring lines. Optimal choice of filter widths and exposure time ratios.
Calibration different. Use emission line sources of known brightness. Need to know filter profile if objects are at different redshifts. Alternatively use stars to get integral under filter, and use ratio of line throughput to total throughput to get flux.
Surface photometry vs. integrated photometry.