Niobium is a reference dosimeter for neutron damage survey of nuclear reactor vessels. This characterization is based on the activity of $^{93m}$Nb resulting from the $^{93}$Nb(n,n') activation reaction. $^{93m}$Nb is a metastable state of $^{93}$Nb that de-excites towards the ground state by a highly converted isomeric transition. This decay mainly results in the emission of niobium K X-rays which are used to determine the activity of $^{93m}$Nb, either by direct measurement using X-ray spectrometry or by liquid scintillation after the sample dissolution. The first method does not require any sample preparation, but fluorescence effects may perturb the measurement. Indeed, niobium foils may contain impurities that are activated during irradiation. Some of these radioactive impurities, in particular $^{182}$Ta, and other niobium isotopes ($^{94}$Nb, $^{95}$Nb and $^{92m}$Nb) are still present during the measurements and disturb the X-ray spectra. The fluorescence effect generates additional niobium X-ray emissions to those due only to the decay of $^{93m}$Nb, resulting in an overestimation of the dosimeter activity. Most activated impurities have a much shorter half-life than $^{93m}$Nb and the current method is to allow the dosimeters to decay for several months before measuring them to obtain their activity. However, in some circumstances, results are expected much earlier, not long after the end of irradiation. It is therefore necessary to assess the influence of the fluorescence effects to provide accurate $^{93m}$Nb activity values. Corrective factors for fluorescence were established by Monte Carlo simulation using the GEANT4 code. This required simulating the complete decay scheme of each radionuclide, taking into account interactions of the emitted particles with matter and atomic relaxation processes. The radioactive impurities were distributed isotropically within the dosimeter and the simulation provided the number of exiting niobium K X-rays as well as the information on the process responsible for the niobium ionization. With this information, it was shown that electrons are mainly responsible for the fluorescence effects in niobium dosimeters. These calculated fluorescence corrective factors were validated by an experimental approach: nine niobium dosimeters with different Ta levels were activated in MARIA reactor (National Centre for Nuclear Research, Poland) in the framework of a round robin initiated by the European Working Group on Reactor Dosimetry (EWGRD). These dosimeters were measured six times in the year following the end of irradiation using gamma- and X-ray spectrometry to determine the impurities activity and the Nb X-rays emission rates respectively. These rates were corrected for the impurities contribution using the calculated corrections before deriving the $^{93m}$Nb activity at the reference date. With the applied corrections, the standard deviation between the six measurements is reduced to about 1 %.