Probing the transition dipoles responsible for the luminescence of a nano-emitter is essential to understand its physical properties, its interactions with its environment and its potential applications. Various methods in photoluminescence microscopy, based on polarimetry or Fourier imaging, have been developed to measure an emitter's dipole properties : number of radiating dipoles, oscillator strength ratio between them, their orientation etc. In this paper, we model the most used of these protocols and show that their sensitivity depends crucially on the experimental conditions : substrate material, presence of another lower or upper layer, objective numerical aperture. We develop guidelines to optimize the measurement sensitivity by tailoring the experimental conditions, depending on the type of protocol used and the dipole property to be measured.