Pattern matching tool of normalized cross correlation operator (NCCO) computation C*(x) for ideal images can be determined by analytical expression Z(x). Compared with C*(x), C(x) of NCCO matching between images containing noise shows a little difference. But there is a property of zero similarity invariance (ZSI). A method of image resolution calibration in microscopic vision was proposed. Firstly, similarity functions of rectangular feature in x direction Z(x) were derived and proved. Then C(x)=0 was solved as an approximation of solution of Z(x). The detailed ZSI based calibrating algorithm was illuminated. The results of simulating image experiments show that the calibration accuracy of the algorithm can reach 0.1～0.2 pixel. Finally, the practical experiments were conducted on a hybrid microassembly workcell. Experimental results show that the calibration precision of the algorithm can reach 0.08 pixel. The ZSI based calibrating algorithm is a method with practical value and subpixel precision.