Temporal Compressive Complex Amplitude Imaging Based on Double Random Phase Encoding

Abstract

Snapshot temporal compressive imaging offers a potent method for capturing high-dimensional spatiotemporal information from a superimposed 2D image of a dynamic scene. However, despite its notable bandwidth-saving capability, simultaneous acquisition of spatiotemporal intensity and phase information remains challenging due to detector phase insensitivity. To address this issue, a novel temporal compressive complex amplitude imaging (TC-CAI) method based on double random phase encoding is proposed here. Within TC-CAI, the target scene undergoes sequential modulation by a static phase mask in spatial domain, followed by spatial encoding by an ultrahigh-speed-switchable phase mask in spatial frequency domain after the Fourier transform. Adjacently, the scene is inversely Fourier transformed and integrally exposed onto a planar detector. Ultimately, the complex amplitude information, sensitive to both amplitude and phase, can be faithfully reconstructed over time using a plug-and-play-based deep image prior algorithm. The feasibility, robustness, and superiority over intensity-based methods of TC-CAI are demonstrated through simulation. This approach is expected to pave the way for real-time multidimensional temporal imaging.

Keywords: Snapshot compressive imaging, temporal imaging, complex-amplitude imaging, compressive sensing, Image Reconstruction, deep image prior