Many times researcher like us go to SEM measurements to study surface microstructure and end with totally unpredictable or irregular shapes. Everytime grain structure can not be uniform (Just like circle, cube, bars) and depends on the methods one has used for the preparation of the samples.
Personal Opinion:
Sometimes SEM images, If not properly recorded may lead to the wrong interpretation and do not add any scientific correlation to any other property.
It is better to perform TEM and as suggested by the other RG members, XRD and RAMAN spectroscopic measurements.
Suggestion:
You can try for 100K X or better resolution for SEM. That may give you better idea.