Deep learning assisted state space method for phase derivative estimation in digital holographic interferometry

Abstract

In digital holographic interferometry, the measurement of derivatives of the interference phase plays a crucial role in deformation testing since the displacement derivatives corresponding to a deformed object are directly related to the phase derivatives. In this work, we propose a recurrent neural network-assisted state space method for the reliable estimation of phase derivatives. The proposed method offers high robustness against severe noise and corrupted fringe data regions, and its performance is validated via numerical simulations. We also corroborate the practical applicability of the proposed method by analyzing experimental data corresponding to deformed test objects in digital holographic interferometry.