Hyperspectral compressive wavefront sensing

• DOI: 10.1017/hpl.2022.35
• Publication Date: 2023-03-21T01:50:35Z
• Authors: Sunny Howard; Jannik Esslinger; Robin H. W. Wang; Peter Norreys; Andreas Döpp
• Keywords: Snapshot (computer storage)

Abstract Presented is a novel way to combine snapshot compressive imaging and lateral shearing interferometry in order capture the spatio-spectral phase of an ultrashort laser pulse single shot. A deep unrolling algorithm utilized for reconstruction due its parameter efficiency superior speed relative other methods, potentially allowing online reconstruction. The algorithm’s regularization term represented using neural network with 3D convolutional layers exploit correlations that exist wavefronts. Compressed sensing not typically applied modulated signals, but we demonstrate success here. Furthermore, train predict wavefronts from interferogram terms Zernike polynomials, which again increases our technique without sacrificing fidelity. This method supported simulation-based results. While example interferometry, methods presented here are generally applicable wide range including Shack–Hartmann-type sensors. results may be interest beyond context wavefront characterization, within quantitative imaging.