The observational signature of supernova remnants (SNRs) is very complex, in terms of both their geometrical shape and their spectral properties, dominated by non-thermal synchrotron and inverse-Compton scattering. We propose a post-processing method to analyse the broad-band emission of SNRs based on three-dimensional hydrodynamical simulations. From the hydrodynamical data, we estimate the distribution of non-thermal electrons accelerated at the shock wave and follow the subsequent evolution as they lose or gain energy by adiabatic expansion or compression and emit energy by radiation. As a first test case, we use a simulation of a bipolar supernova expanding into a cloudy medium. We find that our method qualitatively reproduces the main observational features of typical SNRs and produces fluxes that agree with observations to within a factor of a few. allowing for further use in more extended sets of models.