By P. Hariharan
Publisher: Cambridge University Press
Print Publication Year: 2002
Online Publication Date:July 2010
Chapter DOI: http://dx.doi.org/10.1017/CBO9780511755569.016
Because the intensity in two-beam interference fringes varies sinusoidally with the phase difference, it is difficult to locate the fringe maxima or minima, in a photograph of the interference pattern, to better than a tenth of the fringe spacing. In addition, when the number of fringes is small and they are unequally spaced, errors are introduced by the need for nonlinear interpolation to determine the fractional fringe order at any point.
One way to obtain higher accuracy is by using a CCD camera interfaced with a computer to sample and store the values of the intensity in the interference pattern at an array of points. These values can then be digitized and processed, using a number of techniques, to obtain the fractional fringe order at these points [Robinson & Reid, 1993]. Preprocessing is usually necessary to reduce speckle noise as well as to correct for local variations in image brightness.
An additional tilt introduced in one of the beams (say, along the x direction) generates background fringes corresponding to a spatial carrier frequency. These fringes are modulated by the additional phase difference between the beams due to the changes in the object [Takeda, Ina & Kobayashi, 1982]. If the spatial carrier frequency is sufficiently high, the Fourier transform of the intensity distribution in the interference pattern can be processed to obtain the phase difference.