Fringe pattern recognition and interpolation using nonlinear regression analysis

Abstract

Least-square error criteria are used to fit 1-D interference fringe pattern irradiance data to a physically meaningful function of the form I(x) = B(x) + E(x) cos[P(x)], where B(x), E(x), and P(x) are low-order polynomials. This procedure is intended to complement digital fringe recognition by providing a method for smoothing and interpolating among fringe position data when the number of fringes is small, there are more than ten irradiance measurements per fringe, and accurate phase values are needed at arbitrary locations in the field.