Optical imaging systems are fundamentally constrained by a trade-off between resolution and field of view imposed by finite aperture size. Synthetic aperture imaging overcomes this trade-off by coherently combining multiple measurements to synthesize a larger effective aperture. Fourier ptychography (FP) has emerged as a prominent synthetic aperture imaging technique that reconstructs high-resolution, wide field-of-view images of a complex sample by computationally fusing intensity measurements captured such that they sample distinct regions of the sample’s Fourier space. While ptychography has seen widespread success in microscopy, its application to broader imaging domains has been limited by long image acquisition times, precise calibration requirements, and large computational overhead.

This proposal directly addresses these limitations to enable FP in macroscopic settings such as astronomical imaging. We first show that, by leveraging natural object motion (e.g., the orbital motion of celestial bodies), we can effectively extract frequency information about the target wavefront and recover high-resolution phase and amplitude images of the sample. Furthermore, we introduce a novel implementation of the FP imaging forward model that significantly reduces the computational burden of the reconstruction process.
In particular, we compactly represent the sample with a discrete set of 2D Gaussians and exploit Gaussian properties to eliminate expensive Fourier transform computations during reconstruction. These works advance the versatility and robustness of FP to enable high-quality synthetic aperture imaging across a variety of real-world imaging systems.