a) there won't be many (compared to the number of Zn atoms on lattice sites) and
b) ZnO usually has relatively large amounts of O vacancies (so the Zn/O ratio is usually a little larger than 1 anyway)
The best way would be to look for a characteristic signature in Raman, IR or HREELS spectra, for which you would need firstly a model calculation by DFT or ab initio and, second, an experimental calibration standard.
Quantitative measurement of Zn interstitials are not possible as far as I know. Oxygen vacancies, however can be estimated through iodometric titration. In ZnO, the oxygen vacancies are dominant defect type. For comparative studies suggested by Mr. Thomas Walther, you would need measurements from perfect crystal as control group. DFT or ab initio based calculations for defect calculations will need very high computational power as well as time. Raman, IR and UV Vis spectroscopies can give you a faint idea about different defects, but I doubt if you can quantify them.