[en] The technique of intensity interferometry was largely applied to pairs of bosons produced in heavy ion collisions to study the properties of their source. Recently this technique was applied also to photons which can be considered 'natural' probes in interferometry. The analysis of the results of two experiments, namely Kr + Mi at 60 MeV/N and Ta + Au at 40 MeV/N carried out with the multidetector TAPS at GANIL has shown the complexity of the space-time characteristic of the photon source. The standard hypothesis describing the production of high energy protons (E_γ > 25 MeV) as starting from p-n Bremsstrahlung exclusively in the initial superposition of the two nuclei was rejected. Actually the typical form of the correlation function in which the correlation width corresponds to the inverse of the source size, is not satisfied by any of the two systems. Only, by the taking into account in the BUU calculations the photons produced later bring near the calculations and the data. This late production could originate in the recompression of the di-nuclei system. In analogy with previous application of this technique to stellar interferometry we have studied the structure of the photon source by Monte-Carlo calculations of the correlation function. For the simple case of a binary source the correlation function is dependent on the two source distributions, relative intensity and the space-time separation of the two sources. The results of this calculations evidence the sensitivity of the photon interferometry to different reaction mechanisms by the magnitude and also the shape of the correlation function. The best agreement with the data is obtained when the two nuclear fragments emit simultaneously the photons at a moment subsequent to the reaction moment