Reconstruction in off-axis digital holography based on hybrid clustering and the fractional Fourier transform

Abstract

The fractional Fourier transform (FRFT) introduces fractional-order parameters and provides greater flexibility and adaptability than the conventional Fourier transform, thus offering new possibilities for reconstruction in digital holography. This paper proposes a reconstruction method in off-axis digital holography based on hybrid clustering and FRFT. First, a hybrid clustering algorithm combining K-means and a Gaussian mixture model was used to accurately classify the spectrum information of holograms, effectively suppressing zero-order spectrum interference. Then the rotation additivity property of the FRFT, combined with a particle swarm optimization algorithm, was used to precisely estimate the optimal fractional-order parameters of the hologram. Finally, high-quality amplitude and phase reconstructions were achieved using the FRFT reconstruction of hologram. Simulation and experimental results demonstrated that the proposed method significantly improved the reconstruction quality of holograms. The findings of this study provide an effective solution for enhancing off-axis digital holography imaging quality, with important implications for the development of precision microstructure detection and high-resolution biological imaging technology.