It's rather unlikely you can calculate this from plan-view observation. If you have a multi-granular structure with hexagonal grains and assume these are in fact hexagonal columns you can only fill 3D space completely with certain combinations of diameters, however, your limited SEM resolution and the fact that some grain boundaries may be filled with amorphous material mean you cannot directly see what happens between the grains. For this you need cross-sections and a method like TEM that has (near) atomic spatial resolution, and then correlate this with electrical measurements.

try to use SPIP software on you SEM image( better to have AFM(3D) image) and compare them with XRD result

chose particle and pore analysis

http://www.imagemet.com/

Thank you so much to Thomas Walther and L. Slewa for guide me.