Iterative orthogonal normalization algorithm for improving phase retrieval accuracy

Abstract

The reliance on approximation conditions in conventional orthogonal normalization methods often restricts their accuracy and range of applicability in phase extraction. Consequently, this paper proposes an iterative orthogonal normalization algorithm (IONA) designed specifically for achieving accurate phase retrieval. Unlike traditional orthogonal decomposition methods, this method does not use orthogonal approximation conditions but obtains the correct phase distribution by solving the iterative equation of the phase, transforming the phase solving problem into a loss function minimization problem of the phase distribution. This approach breaks through the limitations of traditional orthogonal methods and can accurately retrieve the phase from the interferograms with less than one stripe. In addition to outperforming existing phase-shifting techniques in accuracy and computational efficiency, the introduced IONA not only negates the requirement for specific fringe counts and configurations but also boasts remarkable stability. Simulation analysis and experimental results indicate that it is a robust and stable solution for phase extraction from fringe patterns with unknown phase shifts.