Multiple signal classification technique for phase estimation from a fringe pattern

The paper introduces a multiple signal classification technique based method for fringe analysis. In the proposed method, the phase of a fringe pattern is locally approximated as a polynomial. The polynomial phase signal is then transformed to obtain signals comprising of only even- or odd-order polynomial coefficients. Subsequently, covariance matrix formulation is applied, and the two sets of coefficients are jointly estimated from the noise subspace of the covariance matrix using the multiple signal classification technique. The method allows simultaneous estimation of multiple coefficients and provides phase without the requirement of complex unwrapping algorithms. The effectiveness of the proposed method is validated through numerical simulation.