Abstract: The interferometer system for transverse beam size measurement was one of the important parts of storage ring diagnostics in Shanghai Synchrotron Radiation Facility (SSRF).The average of multiple lines and curve fitting method was adopted for raw image data processing now, but there were big random errors and system errors when the CCD is not aligned with the double slits. The article discusses the feasibility and the advantages about using primary component analyzing (PCA) method to process the raw image data from the interferometer system. Using MATLAB simulation and two experiments with random noise and rotated CCD, comparing the results errors between PCA raw image data processing and normal average of multiple lines method, it is concluded that the noise could be reduced and the small angle between CCD and double slits could be found out and corrected using PCA methods.